Talk:Artificial intelligence/Archive 10

Evolutionary computation navigation box template
Please view the new evolutionary computation navigation box template: Template: Evolutionary computation. Feel free to help organize and add topics, and place the template at the end of listed articles. Thanks. TheProfessor (talk) 21:45, 28 December 2014 (UTC)

untitled
I, Miguelsnchz723 (talk) 06:01, 3 March 2015 (UTC) am a Business major at CSU Channel Islands and I plan to work on editing this article for my Ethics of the Free World class.

pending edit request
diff

In Applied Behavior Analysis, the science has been democratized into artificially intelligent and intuitive code, allowing a lay person to go online and deliver self-directed clinical treatment in the absence of a clinician based on the positive behavior support principles set by the Behavior Analyst Certification Board. http://autism-u.com - http://medicinex.stanford.edu/conf/submission/view/359 The peer reviewed concept was presented at the annual international conference of the Association for Behavior Analysis (2013) by Deborah Lee Safko, BCBA in an effort to demonstrate the benefits of using artificial intelligence to change the treatment rate for all intellectual disabilities. https://www.abainternational.org/media/67702/ABAImeridaProgram.pdf


 * As I was unsure about this edit I have opened a discussion here, my main issue was the notability and the primary source of the detail, nothing really explained why it was worth reporting it here or whose opinion it was. Please discuss. Govindaharihari (talk) 04:08, 28 November 2015 (UTC)
 * I think you are right -- the sources do not meet the criteria in WP:RS and so this material should not be added to the article. Looie496 (talk) 17:22, 28 November 2015 (UTC)


 * The peer reviewed concept was presented at the annual international conference of the Association for Behavior Analysis (2013) by Deborah Lee Safko, BCBA in an effort to demonstrate the benefits of using artificial intelligence to change the treatment rate for all intellectual disabilities. https://www.abainternational.org/media/67702/ABAImeridaProgram.pdf AI has now been iterated in a manner that allows access to treatment when none would otherwise be available.  In a blind test, a parent can produce the same reports as a clinician, artificially.  https://www.google.com/webhp?sourceid=chrome-instant&rlz=1C1ASUT_enUS515US515&ion=1&espv=2&ie=UTF-8#q=bipartisan+ABA+autism-u A bi-partisan initiative has now been adopted by Senator Marco Rubio and Senator Bill Nelson to recognize the AUTISM-U software as a "provider" by CMS, a designation historically limited to human beings (contact the Senators' offices for verification).  If successful, this initiative would mark a watershed moment in AI, whereby its deployment can change the current treatment rate of the primary treatment for autism, ADHD, etc. from less than 1% to 100%, and potentially save 600 Billion dollars in the projected cost of autism (projected to cost 1 Trillion by 2025, before the new prevalence rates were reported at 1:45). InfoPurist (talk) 15:27, 28 November 2015 (UTC)InfoPurist
 * What part of wiki guidelines does this detail disrupt? Govindaharihari (talk) 20:55, 28 November 2015 (UTC)

First, the peer reviewed sources are within the Wikipedia guidelines. Contrary to the assertion that the citation is an abstract, the content guidelines require precisely that, and to avoid lengthy full text primary sources. https://en.wikipedia.org/wiki/Wikipedia:Do_not_include_the_full_text_of_lengthy_primary_sources

Second, the policy on sourcing is verifiability. The verifiable sources are reliable (Stanford University, Behavior Analyst Certification Board, Florida Association of Behavior Analysts, Association for Behavior Analysis International, and Centers for Medicare and Medicaid Services (CMS) are demonstrable leaders in academics, behavioral health, healthcare, and technology, and Autism-U is sourced as the creator of the work). According to Wikipedia's definition of a source Definition of a source The word "source" when citing sources on Wikipedia has three related meanings: The piece of work itself (the article, book) The creator of the work (the writer, journalist) The publisher of the work (for example, Random House or Cambridge University Press) Any of the three can affect reliability. Reliable sources may be published materials with a reliable publication process, authors who are regarded as authoritative in relation to the subject, or both. These qualifications should be demonstrable to other people.

Third, the company Autism-U is inherently notable according to the terms for notability, as it is the focus of a bipartisan initiative to recognize its software as a provider by CMS, a designation historically reserved to humans, and would represent a watershed moment in artificial intelligence. Arbitrary standards should not be used to create a bias favoring larger organizations or their products, when the information verified and sourced includes the use of artificial intelligence to increase access to behavioral health treatment on a scale never before possible. Wikipedia is supposed to include updated information. Referring to sources in 2007 is relegating Wikipedia users to a historic, and not current, event when the advances in artificial intelligence, according to Moore's Law, demand we have learned something since 2007 which may be beneficial for Wiki readers to review. The ability to self-manage treatment and crowdsource clinical trial data having been verified since 2013, it is relevant to the field of AI that it be recognized as the catalyst, hence the edits to this section. https://en.wikipedia.org/wiki/Wikipedia:Notability_(organizations_and_companies) The arbitrary removal suggests bias for the reasons stated above.

Finally, this detail disrupts the Wikipedia guidelines that editors suggest maintain the dependency on clinicians for verification of any healthcare treatment data. AI enabled self-directed (DIY) behavioral health treatment in the absence of a clinician. It is ironic that the capacity for AI to supercede the current definitions of what it means to be a healthcare provider is being quashed by the editors of the AI page.InfoPurist (talk) 04:15, 1 December 2015 (UTC)InfoPurist


 * I don't feel that the sources are strong enough to justify including this material, and I don't feel that the material has broad enough importance to justify including it in an encyclopedia entry about AI. Looie496 (talk) 12:58, 1 December 2015 (UTC)


 * Agreed. The issue is not WP:notability, or WP:Verifiability. The issue is WP:Relevance. This is the introductory article on artificial intelligence, and artificial intelligence is very large field. I would look for a place in a more specific article. Look through the articles on applications of AI (is there an article with applications of AI in health care?). Or, you could look at the techniques this company is using and place the content down there.   CharlesGillingham (talk) 20:04, 2 December 2015 (UTC)

Section 2.4 "performance at chess is super-human and nearing strong super-human"
Is the quote "performance at chess is super-human and nearing strong super-human" really accurate? If "strong super human" is to mean "better than all humans", I believe it is safe to say computers are strong super human. Stockfish 7, the best chess engine currently, is rated at 3341 ELO, whereas the best human chess player currently, Magnus Carlsen, has an ELO of 2844 --Mornarben (talk) 17:50, 26 February 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just added archive links to 1 one external link on Artificial intelligence. Please take a moment to review my edit. If necessary, add after the link to keep me from modifying it. Alternatively, you can add to keep me off the page altogether. I made the following changes:
 * Added archive http://web.archive.org/web/20090629065702/http://www.shef.ac.uk:80/marcoms/eview/articles58/robot.html to http://www.shef.ac.uk/marcoms/eview/articles58/robot.html

When you have finished reviewing my changes, please set the checked parameter below to true to let others know.

Cheers.—cyberbot II  Talk to my owner :Online 07:02, 27 February 2016 (UTC)

Opinion Based
I just read this article and quite frankly it reads like a list of opinions on the subject. Unfortunately most of these opinions are entirely subjective, including the ones from Alan Turing and many are from people who are not actually researching AI in a professional manner (like Penrose). There has to be a better way to talk about AI than this and I would suggest to separate this article, which is rather philosophical/historical in nature (at best), from an actual technical article about AI proper. I would help with this effort, if I could, but I am a physicist and not a AI researcher and I do feel strongly that an active AI researcher needs to rewrite this in its entirety ASAP. The community may want to try to actively recruit an expert if nobody with a sufficient background has volunteered, yet. 87.229.87.235 (talk) 14:34, 24 July 2015 (UTC)
 * Disagree -- read the "tools" and "approaches" sections. CharlesGillingham (talk) 20:25, 2 December 2015 (UTC)
 * I agree that it is opinion based. Read this line in History section "In the 1990s and early 21st century, AI achieved its greatest successes, albeit somewhat behind the scenes." - there is no source to back this up. Trt94
 * The source for those two sentence is in footnote 12. The first source is the most widely used academic AI textbook, which makes this point explicitly. --- CharlesGillingham (talk) 17:34, 7 March 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just added archive links to 2 one external links on Artificial intelligence. Please take a moment to review my edit. If necessary, add after the link to keep me from modifying it. Alternatively, you can add to keep me off the page altogether. I made the following changes:
 * Added archive http://web.archive.org/web/20100724025845/http://www.gadgetrivia.com:80/8164-best_robot_international_robot_exhibition to http://www.gadgetrivia.com/8164-best_robot_international_robot_exhibition
 * Added archive http://web.archive.org/web/20131004222242/http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4895715&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel5%2F4563672%2F5038478%2F04895715.pdf%3Farnumber%3D4895715 to http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=4895715&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel5%2F4563672%2F5038478%2F04895715.pdf%3Farnumber%3D4895715

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.—cyberbot II  Talk to my owner :Online 08:08, 10 March 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just added archive links to 2 one external links on Artificial intelligence. Please take a moment to review my edit. If necessary, add after the link to keep me from modifying it. Alternatively, you can add to keep me off the page altogether. I made the following changes:
 * Added archive http://web.archive.org/web/20091008184132/http://lis.epfl.ch:80/resources/podcast/2007/11/gerald-edelman-neural-darwinism-and.html to http://lis.epfl.ch/resources/podcast/2007/11/gerald-edelman-neural-darwinism-and.html
 * Added archive http://web.archive.org/web/20081007162810/http://www.rci.rutgers.edu:80/~cfs/472_html/AI_SEARCH/PSS/PSSH4.html to http://www.rci.rutgers.edu/~cfs/472_html/AI_SEARCH/PSS/PSSH4.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.—cyberbot II  Talk to my owner :Online 01:47, 27 March 2016 (UTC)

Should unqualified opinions be included?
The article currently cites Stephen Hawking. With the greatest of respect to Stephen Hawking's contributions to physics, could someone please clarify what qualifies his opinion on AI? Why should it appear here and is it just his celebrity status that adds weight? Would it be reasonable to cite AI dons in theoretical physics articles? Perhaps this issue has already been raised before and I missed it, in which case I apologise in advance. pgr94 (talk) 16:40, 12 April 2016 (UTC)

re-write intro paragraph
The bulk of the article is good. But the introductory paragraph, defining AI, is just a bunch of sentences which use the words artificial and intelligence in them, hardly a useful layman's definition of the subject (which it should be). How about using words like cognition? Pb8bije6a7b6a3w (talk) 13:08, 17 May 2016 (UTC)

trendy AI definition
I don't agree with the idea in the intro that AI is something that mimics intelligence, and we've already accomplished it. True AI is something that has cognitive ability. Your intro's definition can be accomplished by any decent programmer and a bunch of IF-THEN statements. But no one can write the program that asks itself who it is and what it's doing.Pb8bije6a7b6a3w (talk) 17:44, 30 May 2016 (UTC)
 * Nice change in intro! Pb8bije6a7b6a3w (talk) 23:25, 1 June 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 1 one external link on Artificial intelligence. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20070926023314/http://www.aaai.org/AITopics/assets/PDF/AIMag26-04-016.pdf to http://www.aaai.org/AITopics/assets/PDF/AIMag26-04-016.pdf

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 14:19, 11 September 2016 (UTC)

Good Article
This wikiarticle is a very good article showing a lot of great information and sources. I think the by including the chess example as well as the self-driving cars show the timeline of research that has been put into artificial intelligence. Along with this, I think that there is a lack of informations as to explaining what exactly the point of artificial intelligence is. The article goes very well into extensive research as to what has been created and various projects, but the why is lacking at the beginning of the article especially. I feel as if this article is very neutral in what it is trying to explain about the research and progression of artificial research. I looked at the source list and did find that there was some repeated authors that were referenced. With this being said, on the reference list I did not see the same author sited more then twice. With this being said, I feel that with as long of a source list as this, I thin it is bound to happen that a repetition will occur. The large source list does show that there is very reliable and a large number of authors and sources that cross-reference each other proving this article to be reliable. Going through some of the authors on the source list such as Stuart J. Russell I researched a bit about him and the first two sets of information that came up was he was an author to two books on Artificial Intelligence. The first being Artificial Intelligence: A Modern Approach and there was multiple editions on this book from him. I felt that this strengthened the articles reliability with just one of the many authors having multiple books out on the topic of Artificial Intelligence. I feel that the information with the self-driving car brings the article up to date because it is a current scientific topic with Uber and the Tesla company. ''Charvey1597 (talk) 00:53, 16 September 2016 (UTC)

Excellent Article
A truly excellent article, strange and beautiful like AI itself. Scope creep (talk) 10:19, 11 October 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 4 one external links on Artificial intelligence. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20120611115223/http://www.thenewatlantis.com:80/publications/artificial-intelligence-and-human-nature to http://www.thenewatlantis.com/publications/artificial-intelligence-and-human-nature
 * Added archive https://web.archive.org/web/20130118095844/http://www.csjournals.com/IJITKM/PDF%204-1/35.Sandeep%20Rajani.pdf to http://www.csjournals.com/IJITKM/PDF%204-1/35.Sandeep%20Rajani.pdf
 * Added archive https://web.archive.org/web/20100118120034/http://www.bbsonline.org:80/documents/a/00/00/04/84/bbs00000484-00/bbs.searle2.html to http://www.bbsonline.org/documents/a/00/00/04/84/bbs00000484-00/bbs.searle2.html
 * Added archive https://web.archive.org/web/20090325194051/http://www.patternrecognition.co.za/publications/cvdwalt_data_characteristics_classifiers.pdf to http://www.patternrecognition.co.za/publications/cvdwalt_data_characteristics_classifiers.pdf

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 02:45, 19 October 2016 (UTC)

IAH206 bibliography for proposed edits
Here is a bibliography of some sources that could be used to update and contribute to this article.

Bass, Dina. "Microsoft Develops AI to Help Cancer Doctors Find the Right Treatments." Bloomberg. N.p., n.d. Web. 22 Oct. 2016

Clark, Jack. "Why 2015 Was a Breakthrough Year in Artificial Intelligence." Bloomberg. N.p., n.d. Web. 22 Oct. 2016.

McFarland, Matt. "Google's artificial intelligence breakthrough may have a huge impact on self-driving cars and much more." Washington Post 25 Feb. 2015. Infotrac Newsstand. Web. 24 Oct. 2016

"Partnership on Artificial Intelligence to Benefit People and Society."Partnership on Artificial Intelligence to Benefit People and Society. N.p., n.d. Web. 24 Oct. 2016.

“Programming safety into self-driving cars.” National Science Foundation. N.p., 2 Feb. 2015. Web. 24 Oct. 2016.

Rowen, Norma. “The Making of Frankenstein’s Monster: Post-Golem, Pre-Robot.” Pennsylvania Electronic Edition. N.p., n.d. Web. 23 Oct. 2016

Sage, Alexandria. "Self-driving Car Startup Drive.ai Names Former GM Exec to Board." CNBC. N.p., n.d. Web. 22 Oct. 2016.

Ozogjess (talk) 00:21, 25 October 2016 (UTC)Ozogjess


 * In general a source like "Microsoft Develops AI to Help Cancer Doctors Find the Right Treatments" saying "here's yet another way that some organization has used AI" isn't likely to add much to the article, because there are already far too many to itemize here. Something like "Why 2015 Was a Breakthrough Year in Artificial Intelligence" is closer to being a useful source, but its focus on 2015 is kind of narrow. Rolf H Nelson (talk) 05:56, 8 November 2016 (UTC)

IAH206 proposal for edits
We propose to update and include more recent information regarding artificial intelligence. We want to focus mostly on newer advancements and technology involving artificial intelligence from approximately 2000 to the present. We will incorporate information about how artificial intelligence is being used in the medical industry, about how it is being integrated into the automotive industry and self-driving cars, and about the future and advancements of artificial intelligence. We will bring information to the article regarding specific fiction novels that display the artificial intelligence theme. Lastly, we will add information about the latest genetic cloning techniques, how scientists implement genetic cloning with the use of artificial intelligence, and the moral implications behind it.

Ozogjess (talk) 22:34, 31 October 2016 (UTC)Ozogjess
 * "In fiction" is already broken off into its own separate article. I'd consider moral implications of cloning to be off-topic in the current article. Rolf H Nelson (talk) 05:56, 8 November 2016 (UTC)

Suggestions for improvement(IAH)
·     In the history section: o  The subsection titled 2015: A breakthrough just seems odd. The format and blatancy of the bold font title looks ill-placed. ·     Run-on sentence in section Research, sub-section Knowledge Representation. ·     The whole Research sections seems excessively long. I understand that Artificial Intelligence is a very in depth topic, but I feel that the Research section constitutes a majority of the article. The sub-section Tools could potentially be separated from the rest of the section, and maybe more sub-sections could be moved out of the section. o  A more critical focus on this part of the article is suggested to trim and cut down this section. The whole Wikipedia article itself is quite lengthy. Also, this article is very long and while an important section that I feel you could expand on, Long term goals, only has one line. Whereas, sections that I believe the reader would be less interested in such as, classifiers and statistical learning methods nearly has 250 words. Therefore, I think it would benefit your article to expand on sections such as long term goals of AI so that the reader of the article is satisfied with the articles information. In addition, the section 2015 a breakthrough, could have been underrepresented because there were enough facts there that could be drawn from at least another source. Welshtor (talk) 04:47, 22 November 2016 (UTC)MattClark (talk) 04:32, 22 November 2016 (UTC) Matthew Clark, November 21, 2016

Standalone healthcare page
I would like to propose a standalone page for 'Artificial intelligence in healthcare'. There are currently standalone pages for AI and law, AI marketing and AI video surveillance, so an article for AI in healthcare seems like a natural addition. There is an extensive body of research on the topic to cite. Would love to hear your thoughts? Laurenmcq (talk) 01:06, 29 November 2016 (UTC)
 * If you feel like doing the work of putting that page together, sure, absolutely, go for it. Looie496 (talk) 15:08, 29 November 2016 (UTC)

Super-human?
The categorisation in the section Evaluating progress is very questionable. It includes "Super-human: performs better than most humans". Yet the conventional definition of superhuman qualities are qualities that exceed those found in humans. A horse is superhuman in its running ability because no human can keep up. A green iguana can probably travel faster than 80% of humans but at 22mph is not superhuman in speed. In the section it has been necessary to invent a category "strong super-human" to make up for this abuse of English.

The reference given for this dates from 2011, does not appear to be by an expert and was probably copied straight from Wikipedia. The list has been there for many years and its 4 categories may be useful. But unless someone can produce some justification for the names I think they should be changed. I've already changed the list in the Progress in artificial intelligence article and could change this to align with that unless someone has a better suggestion. Chris55 (talk) 19:09, 5 February 2017 (UTC)
 * No original research on Wikipedia, please. Here is a source:

For instance, one way of evaluating AI progress is to look at a task and check in which category an AI system is placed: optimal if no other system can perform better, strong super-human if it performs better than all humans, super-human if it performs better than most humans, par-human if it performs similarly to most humans, and sub-human if it performs worse than most humans (Rajani 2011). Note that this approach does not imply that the task is necessarily evaluated with a human-discriminative approach. Having these categories in mind, we can see how AI has scaled up for many tasks, even before AI had a name.
 * from page 27. Hernández-Orallo (2016) Evaluation in artificial intelligence: from task-oriented to ability-oriented measurement. Artificial intellence review. pgr94 (talk) 21:06, 5 February 2017 (UTC)
 * This discussion was started here pgr94 (talk)
 * Now you are quoting a 2016 reference as a replacement for a 2011 citation of (possibly) a Wikipedia article of something that you added to the article on 18 Jan 2008 without any citation. This is how myths and rumours are spread; I wouldn't grace it with the name of research. And dictionary definitions aren't research either. Chris55 (talk) 23:04, 5 February 2017 (UTC)
 * Chris, I don't understand the contention. These are peer-reviewed scientific articles.  Rajani (2011) appears to be the primary source and Hernández-Orallo (2016) is a secondary source, citing Rajani.  Whether Rajani copied my contribution to Wikipedia is speculation and my understanding is that this has no bearing on the validity of these papers as sources.  If you have better sources for a different categorisation do introduce them.  But introducing the term "high human" as you have done in Progress in artificial intelligence without supplying any sources constitutes original research. pgr94 (talk) 01:43, 6 February 2017 (UTC)
 * Pgr94, let me spell it out for you simply. The statement in Rajani's paper is:
 * The broad classes of outcome for an AI test are: optimal in which it is not possible to perform better, strong super-human which performs better than all humans, super-human which performs better than most human and sub-human which performs worse than most human.
 * Apart from some joining words and a couple of grammatical errors this is identical with the statement that is in the article and that has been there since 2008. It even repeats the hyphens which are rarely used in the word "superhuman". If I was marking this as a student paper, given the checking software routinely used these days, it would probably rank as plagiarism. How else do you think that the editor who added the reference (which replaced a redlined faulty reference) and who is clearly competent but a non-specialist, found it?
 * There are thus two possibilities: either you both copied it from a published source, in which case the onus is on you to produce a proper citation, which you haven't done in more than 8 years, or as is more likely, Rajani copied it from Wikipedia.
 * However my primary objection is still to the misuse of the English word "superhuman" in the definition. You haven't yet convinced me. Chris55 (talk) 09:13, 6 February 2017 (UTC)

Since I don't see a direct way of resolving this issue, I've requested a Peer review of the Progress in artificial intelligence article which expands this section. Chris55 (talk) 12:00, 7 February 2017 (UTC)

Pgr94 said "and my understanding is that this has no bearing on the validity of these papers as sources. " Your understanding is wrong. WP:REFLOOP forbids citing sources that get their facts from Wikipedia. It suggests finding the original source that Wikipedia used, and citing that directly. While I don't think there's any proof that this is a refloop situation, it's very understandable why Chris55 would want a pre-2008 source for a contentious edit made in 2008.ApLundell (talk) 15:41, 7 February 2017 (UTC)

As for the point under discussion, I don't understand why this is a big deal. If that is a commonly used classification scheme, it should be really really easy to find many citations. If it was just used by one or two people, it adds no value to the article and could be easily replaced with any other (less controversial) classification scheme. I personally can't find much support for this strangely worded classification scheme, but I'm not an academic, so I reserve judgement because I may be simply missing important sources. I do agree that "super-human" to mean "better than average human" is not common usage. (Imagine if DC Comics used that definition!) ApLundell (talk) 15:41, 7 February 2017 (UTC)


 * Thanks for your first point; as to whether this is commonly used, I would say as one who's been involved with AI for a very long time that this is the first such scheme I've come across. It's probably original: not a bad thing in itself. The questions are therefore (a) whether a citation is needed (b) whether it's worth while (c) how can one avoid giving the impression that it's more than a presentation technique in the article. Chris55 (talk) 18:33, 7 February 2017 (UTC)
 * If it's not an industry standard, the odd use of the phrase "super-human" makes it more confusing then necessary. In that case, I don't see why we need to define terms at all, the concept of "better than the average human" is pretty easy to understand without inventing jargon. ApLundell (talk) 18:39, 7 February 2017 (UTC)
 * Ok, I've aligned the list of categories with the Progress article and removed the citation. This paragraph is clearly OR and I'm not very happy with it, but it's also useful so I'm not inclined to delete it. Chris55 (talk) 14:20, 10 February 2017 (UTC)
 * I understand the WP:REFLOOP concern but I still don't agree with you introducing a new term "high human" with no sources at all. For me, super-human means better than most humans.  "Most humans" means the majority of humans, which is more than 50%.  "High human" means nothing to me at all.pgr94 (talk) 14:48, 10 February 2017 (UTC)
 * Merriam Webster defines "superhuman" as "exceeding normal human power, size, or capability" . "Normal human power" implies average or median; not best human. pgr94 (talk) 14:51, 10 February 2017 (UTC)
 * That is the second definition of the word. Average and normal are not the same thing. The OED supplies the rather fuller second definition as "In rhetorical or hyperbolical use: beyond that of any ordinary person; extraordinary, exceptional." The first definition in the same dictionary is "Of a person, being, or agent: more than human; having a nature superior to that of an ordinary human." That is the more correct sense: one example it gives is Christ. As I said earlier, by your use, every Mensa member would have superhuman intellect and that is clearly misleading. Chris55 (talk) 19:16, 10 February 2017 (UTC)

Evaluating progress in artificial intelligence
I have removed the unsourced material here as the categorisation falls foul of WP:OR. Personally, I think the categories from the Artificial intelligence review are acceptable despite WP:REFLOOP. Let's wait until a better source is identified. pgr94 (talk) 14:25, 21 February 2017 (UTC)

Heideggerian AI
Hi there, just wondering why there is no section on Heidegerrian AI? I once had the opportunity to talk with Marvin Minsky and he told me that he felt "mainstream AI (of the early 2000's) was brain dead." I don't agree, but over the years I have come to see the difference between "ordinary AI" and Heidegerrian AI. I thought I check on to the latest thoughts on this on the Wikipedia but was a bit surprised to see it isn't even mentioned in the page, other than links to Dreyfus. Please advise. Sol ai (talk) 12:42, 20 March 2017 (UTC)

Length
I argue that this is WP:Summary article of a large field, and that therefor it is okay that it runs a little long. Currently, the article text is at around ten pages, but the article is not 100% complete and needs more illustrations. CharlesGillingham (talk) 18:29, 2 November 2010 (UTC)

Todo: Long-Term Goals
The long term goals page is not comprehensive. It also has no sources. Needs immediate editing Trt94

Todo: Illustration
The article needs a lead illustration and could use more illustrations throughout. CharlesGillingham (talk) 18:29, 2 November 2010 (UTC)
 * Thanks to User:pgr94, the article is 70% illustrated. Almost there. CharlesGillingham (talk) 00:03, 16 June 2011 (UTC)
 * Main illustration doesn't provide an actual example of an Artificial Intelligence, just a robot capable of mimicking human actions in a certain area (Namely, sport) — Preceding unsigned comment added by 86.163.226.52 (talk) 15:37, 4 August 2011 (UTC)
 * I've reverted the addition of a picture of the replica of a fictional AI to the lead. I don't think this fits the focus of this article which is about real-life AI endeavours (see for example comments in ). --Mirokado (talk) 13:15, 25 October 2014 (UTC)
 * While the importance of limiting the article's content to real science cannot be understated, the current lead illustration, an "AI themed mug," seems to undercut the seriousness of the topic more than a picture of HAL 9000 would. ScreamingRobot (talk) 00:24, 21 May 2015 (UTC)

Todo: Applications
The "applications" section does not give a comprehensive overview of the subject. CharlesGillingham (talk) 18:29, 2 November 2010 (UTC)

Todo: Topics covered by major textbooks, but not this article
I can't decide if these are worth describing (in just a couple of sentences) or not. CharlesGillingham (talk) 18:29, 2 November 2010 (UTC)
 * 1) Could use a tiny section on symbolic learning methods, such as explanation based learning, relevance based learning, inductive logic programming, case based reasoning.
 * 2) Could use a tiny section on knowledge representation tools, like semantic nets, frames, scripts etc.
 * 3) Control theory could use a little filling out with other tools used for robotics.
 * 4) Should mention Constraint satisfaction. (Under search). Discussion below, at Talk:Artificial intelligence/Archive 4.
 * 5) Should mention the Frame problem in a footnote at least. CharlesGillingham (talk) 19:52, 3 February 2011 (UTC)

Todo: redlinks and tags

 * 1) Where can we link Belief calculus? Does this include Dempster-Shafer theory (according to R&N)? I think that's more or less deprecated. Does R&N include expectation-maximization algorithm as a kind of belief calculus? I don't think so. Where is this in Wikipedia?
 * 2) There are still several topics with no source: Subjective logic, Game AI, etc. All are tagged in the article. CharlesGillingham (talk) 19:59, 3 February 2011 (UTC)

Cognitive systems?
Curiously, this article is a redirect from "cognitive systems" but does not mention that term anywhere. It would seem that there is still room for a separate, introductory article on Cognitive Systems what does the term mean? What are C.S.? What do they do? Or is this all covered adequately in cognition? Ralohmann (talk) 13:01, 6 April 2016 (UTC)

Thanks to the author(s) from the Tatar Wikipedia participants
Thank you, the author(s) of this article. We translated your article into the Tatar language.--A.Khamidullin (talk) 12:41, 26 March 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 6 external links on Artificial intelligence. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Corrected formatting/usage for http://www.thenewatlantis.com/publications/artificial-intelligence-and-human-nature
 * Corrected formatting/usage for http://www.isd.mel.nist.gov/documents/albus/4DRCS.pdf
 * Corrected formatting/usage for http://www.engagingexperience.com/2006/07/ai50_ai_past_pr.html
 * Added archive https://web.archive.org/web/20130118095844/http://www.csjournals.com/IJITKM/PDF%204-1/35.Sandeep%20Rajani.pdf to http://www.csjournals.com/IJITKM/PDF%204-1/35.Sandeep%20Rajani.pdf
 * Added archive https://web.archive.org/web/20100118120034/http://www.bbsonline.org/documents/a/00/00/04/84/bbs00000484-00/bbs.searle2.html to http://www.bbsonline.org/documents/a/00/00/04/84/bbs00000484-00/bbs.searle2.html
 * Added archive https://web.archive.org/web/20090325194051/http://www.patternrecognition.co.za/publications/cvdwalt_data_characteristics_classifiers.pdf to http://www.patternrecognition.co.za/publications/cvdwalt_data_characteristics_classifiers.pdf

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 23:20, 12 June 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 2 external links on Artificial intelligence. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20141129044821/http://www.aaai.org/Library/Symposia/Fall/fs05-06 to http://www.aaai.org/Library/Symposia/Fall/fs05-06
 * Added archive https://web.archive.org/web/20150713184128/https://www.techcastglobal.com/documents/10193/34869/Consciousness-Sagifinalversion to https://www.techcastglobal.com/documents/10193/34869/Consciousness-Sagifinalversion

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 00:41, 10 July 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Artificial intelligence. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20141112130954/http://www.rethinkrobotics.com/artificial-intelligence-tool-threat/ to http://www.rethinkrobotics.com/artificial-intelligence-tool-threat/

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 14:35, 2 August 2017 (UTC)

this is cool — Preceding unsigned comment added by 66.57.220.146 (talk) 16:02, 22 August 2017 (UTC)

Should the huge increase in interest in AI be mentioned?
Anyone who's kept up with the subject over the last few years will have noticed how much more often AI is now discussed in media and by prominent figures. Very few of these figures ever seem to dispute its current rapid advancement, but rather there seems to be a split between those who warn of its destructive potential and those who focus on how it may help us. Network World quotes Jeff Bezos as believing we're in a golden age of AI, and the hype surrounding it now has been noted. "AI revolution" seems to have entered common usage as well. Bataaf van Oranje (Prinsgezinde) (talk) 11:21, 29 September 2017 (UTC)

Rusty's review
The overview is pretty comprehensive. It hits a lot of major points on Artificial Intelligence. Several links to references provide you with an excellent scope of knowledge on the subject. One improvement could be to address how complex the programming is and to discuss the difference between sentience and intelligence. — Preceding unsigned comment added by 8.41.197.222 (talk) 19:27, 2 October 2017 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Artificial intelligence. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20110628194905/http://www.csse.monash.edu.au/publications/1997/tr-cs97-322-abs.html to http://www.csse.monash.edu.au/publications/1997/tr-cs97-322-abs.html

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 00:02, 21 November 2017 (UTC)

Game design
Why is the template transcluded in the article? Designers can put some AI in their games, but now a days you can put AI in almost all software so that is not a very good reason. Hevesli (talk) 17:32, 2 January 2018 (UTC)


 * Good question. It's probably from the olden days before there was a Artificial intelligence (video games) article.  That template should be moved there. ApLundell (talk) 21:04, 2 January 2018 (UTC)


 * I see that you've moved the template. Good. I've gone ahead and edited the template so that the "Artificial Intelligence" link goes to the game article and not this general article. ApLundell (talk) 20:52, 3 January 2018 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on Artificial intelligence. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20071012025921/http://www.acm.org/class/1998/I.2.html to http://www.acm.org/class/1998/I.2.html

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 06:12, 9 January 2018 (UTC)

"Basics" section
This section is accurate and well written and well cited, and may be needed in some form. However, beyond the first sentence, it only describes statistical approaches to machine learning, and overlooks all the other unsolved problems and deeper issues that the field has wrestled with for the past 60 years. Now, I suppose that one could argue that in 2018, AI just is statistical approaches to machine learning. There are many people today who talk this way, throughout the business world and the vast majority of academia. But, as someone who's involvement in AI predated all of this by a few decades, I'm bothered that the field has narrowed so much on these problems and these solutions. It ignores all the unattempted problems*, the value of alternative approaches, and how very far the field has to go. CharlesGillingham (talk) 01:28, 24 March 2018 (UTC)

\* consider the commonsense knowledge problem, e.g. "If Abraham Lincoln was in Washington where was Abraham Lincoln's left foot?" There is no statistical approach for this. You need knowledge and reasoning, statistical methods will not help. There is no training data: this question and its answer probably doesn't appear anywhere in the vast corpus of world's text, except as an example of the CKP. CharlesGillingham (talk) 01:34, 24 March 2018 (UTC)
 * To the extent the field has narrowed, Wikipedia can't do anything about that; we have to reflect the sources. That said, there's certainly more core concepts to add; I went ahead and added commonsense reasoning and I also made it clearer that there are different approaches. Rolf H Nelson (talk) 22:54, 24 March 2018 (UTC)
 * I've always looked to the AI textbooks to answer these questions. If Russell and Norvig cover it, then it is AI.  CharlesGillingham (talk) 16:54, 26 March 2018 (UTC)

Removed section on Control Theory
Control theory is not a field of AI. I removed it from that section. Ergzay (talk) 07:07, 26 February 2018 (UTC)


 * Control theory is discussed as a tool for AI in Russell and Norvig's 2003 edition of Artificial Intelligence: A Modern Approach, which was the most popular AI textbook when the article was new. I don't know if it is still part of the latest edition of their textbook, or other textbooks which may be more popular now. CharlesGillingham (talk) 20:38, 27 March 2018 (UTC)

WP:Relevance
This article has always been dogged by problems with WP:Relevance. Back in 2007 I tried to fix all these problems empirically with this:


 * Talk:Artificial intelligence/Textbook survey

Of course, this is very out of date. But I just wanted to make editors aware of that this is a problem and that there is a way to fix it empirically, if you are so inclined. CharlesGillingham (talk) 20:54, 27 March 2018 (UTC)
 * Thanks for the pointer, though (as you said) it is out of date (1998-2004) and with the usual caveat that Wikipedia is WP:NOTTEXTBOOK not a textbook, this is a very useful resource. Rolf H Nelson (talk) 04:23, 29 March 2018 (UTC)
 * I don't think WP:NOTTEXTBOOK applies here. That has more to do with how the article is written. The trouble is that there are thousands of topics that can be covered in an article about AI, and we don't want miss any essential ones, and we need to skip everything that isn't essential. On top of that, contributors arrive here with a view of AI shaped by the business world, or by what they read in fiction, or by what their professors in school were working on, or by their own original thoughts about this interesting subject. Russell & Norvig provides a solid example that is more balanced than what you get elsewhere.  CharlesGillingham (talk) 23:45, 30 March 2018 (UTC)
 * True, rather than reference WP:NOTTEXTBOOK I should have said that have a broader audience than a textbook and also have a broader scope than a textbook.Rolf H Nelson (talk) 06:28, 3 April 2018 (UTC)
 * I agree with the "broader audience" part -- which is why we have proportionally larger sections on social history, technical history ("Approaches"), philosophy, speculation. CharlesGillingham (talk)

Missing AI topic - Artificial intelligence used in Software bug detection/quality assurance
I have perused the Wikipedia articles on both "AI" and "Expert systems" and found no reference to the early use of AI in detecting (or predicting) problems or locating potential bugs in application or systems programs. As someone who was instrumental in developing such systems from the 70's through to 1994 for IBM mainframes (IBM/360,370,390,ES9000, etc) and Operating Systems, I find this extraordinary.

Where better to utilise AI techniques in the first instance than in the creation of reliable, quality checked computer programs you might well ask? At the time of developing these systems there was precious little in the way of introspective computing. My software may have actually been the first of its genre and foreshadowed the likes of "Visual Studio" and similar products which of course were produced for a different generation of hardware. My software automatically also prevented buffer overflow, tight loops, macro loops and excessive resource utilisation to name but a few things.

Of course, this lack of knowledge is heightened by the fact that the details of the products that I created (OLIVER & SIMON) have been deleted from Wikipedia over the years through Wikipedia's strict rules (Original research/citations). This is despite the fact that these products were in commercial use with around 600+ large IBM data centres in Europe and elsewhere over an extended period during the 70's, 80's and 90's. Derivatives of these products are still in use in 2018. — Preceding unsigned comment added by 2A00:23C1:4DB2:2400:DC7E:E912:1EA8:AA82 (talk) 10:19, 8 April 2018 (UTC)
 * As already stated in the article, "AI is relevant to any intellectual task. Modern artificial intelligence techniques are pervasive and are too numerous to list here." Regarding work from the 20+ years ago, it might also fall into "Frequently, when a technique reaches mainstream use, it is no longer considered artificial intelligence". If you have a strong source of it as a prominent use of artificial intelligence, Applications of artificial intelligence or (if the application is narrow) Test automation would probably be a good place to start. Rolf H Nelson (talk) 04:42, 10 April 2018 (UTC)


 * The field is truly gigantic and ancient. Entire sub-fields and historical ages involving thousands of people and hundreds of institutions have disappeared beneath the surface.


 * In the 70s and 80s AI-based automatic programming was a very promising direction, and it seemed like we were on our way to bug-free, perfectly efficient computer programs by the turn of the century. However, this all became very unfashionable in the 90s: the lower quality standards of internet programming, agile software development, the giant step backward. So automatic programming is on the back burner for now. Sadly, the total amount of work being done in this vein is genuinely minuscule today (relative to the amount of resources going into deep learning, for example.) So it doesn't have much of a footprint here in Wikipedia today.


 * I'm sure automatic programming will come back into fashion one day. It seems ridiculous to me that we are willing to put up with all these bugs, crashes, and slow running software, when AI was so close to solving this back in the 80s. CharlesGillingham (talk) 23:16, 12 May 2018 (UTC)

AI Languages/Platforms ?
Anyone else think History of software engineering might be useful/helpful enough to be RETURNED TO returned/DUPLICATED(but WP:CWW?) in this article?? --Curious1i (talk) 04:32, 17 May 2018 (UTC)
 * If there's no WP:CONSENSUS on its removal from here we should restore it; the edit where I removed the material is here. It would indeed be useful if we can get a third opinion on it from other page editors. Keep in mind that the article WP:LENGTH of the readable prose in this article is currently around 90kb, which is a little large by wikipedia standards. Rolf H Nelson (talk) 07:34, 17 May 2018 (UTC)


 * I personally don't think that the section you linked to is particularly good. The section in this article has been vetted and re-vetted for relevance, whereas the section you linked only mentions a few randomly selected facts. Just my opinion.  CharlesGillingham (talk) 20:20, 17 May 2018 (UTC)


 * I guess the (simplistic) way I think about it is: AI Languages(/Platforms) WAS a part of this AI article for years (I'm thinking about importance of mainly Lisp & Prolog...), AND, it seems important enough to appear in Template:Artificial intelligence... — Preceding unsigned comment added by Curious1i (talk • contribs) 02:43, 19 May 2018 (UTC)
 * What if we have a paragraph on AI programming trends that links to the new History of software engineering section, which restores mention of lisp and prolog, and could include other information such as a brief sentence about the current talent crunch and that there has been a rapid growth in AI since 2012. Rolf H Nelson (talk) 04:49, 24 May 2018 (UTC)

Edit for Clarification
By the "statistical learning" subsection, the abbreviation "GOFAI" has not be defined in several sections. As many readers jump directly to certain sections, it might help ease of readership to redefine this term every once in a while. Stratovarius03 (talk) 01:52, 21 June 2018 (UTC)
 * There's never a perfect solution for what to do about technical jargon given that most users jump around sections, but I added "traditional" to soften one of the re-introductions. Rolf H Nelson (talk) 18:45, 14 July 2018 (UTC)

Possible Missing Citation
In the overview, it appears a citation should be given for the statement "Many tools are used in AI, including versions of search and mathematical optimization, artificial neural networks, and methods based on statistics, probability and economics. The AI field draws upon computer science, mathematics, psychology, linguistics, philosophy and many others."

No citation is provided for either assertion. Stratovarius03 (talk) 01:41, 21 June 2018 (UTC)


 * The first sentence is a summary of the "tools" section in the article, and citations can be found there. The second sentence is fairly obvious to anyone familiar the field and its history. But you could probably find a citation in the introductory "what is AI" chapter of any AI textbook.  CharlesGillingham (talk) 05:52, 15 July 2018 (UTC)

Semi-protected edit request on 28 July 2018
In the machine ethics section of the article. The sentence, "Recognition of the ethical ramifications of behavior involving machines, as well as recent and potential developments in machine autonomy, necessitate this." Should have a citation. I suggest this one: https://www.yalelawjournal.org/article/machine-testimony

In the section listing applications, add the category "Art."

In the category "Art," include the following description:

Artificial Intelligence has inspired numerous creative applications including its usage to produce visual art. The exhibition "Thinking Machines: Art and Design in the Computer Age, 1959-1989" at MoMA (https://www.moma.org/calendar/exhibitions/3863) provides a good overview of the historical applications of AI for art, architecture, and design. Recent exhibitions showcasing the usage of AI to produce art include the Google-sponsored benefit and auction at the Gray Area Foundation in San Francisco, where artists experimented with the deepdream algorithm (https://www.washingtonpost.com/news/innovations/wp/2016/03/10/googles-psychedelic-paint-brush-raises-the-oldest-question-in-art/?noredirect=on&utm_term=.11818aba2044)and the exhibition "Unhuman: Art in the Age of AI," which took place in Los Angeles and Frankfurt in the fall of 2017.(https://www.statefestival.org/program/2017/unhuman-art-in-the-age-of-ai) (https://www.artsy.net/article/artsy-editorial-hard-painting-made-computer-human). In the spring of 2018, the Association of Computing Machinery dedicated a special magazine issue to the subject of computers and art highlighting the role of machine learning in the arts. (https://dl.acm.org/citation.cfm?id=3204480.3186697) NDPlume (talk) 22:45, 28 July 2018 (UTC)


 * I put it in because it is a good faith edit, not a decision by me that it should stay in. Sincerely, North8000 (talk) 16:37, 29 July 2018 (UTC)

Semi-protected edit request on 13 August 2018
The link for footnote 337 is supposed to point to the Verge article in question, but leads instead to https://en.wikipedia.org/wiki/Template:Cite_web page. Phineas Cage (talk) 10:09, 13 August 2018 (UTC)
 * Edit-undo.svg Undone: This request has been partially undone. I will comment in the section above. O3000 (talk) 11:31, 13 August 2018 (UTC)

Semi-protected edit request on 29 August 2018
Consider changing "unabatedly" to "unabated," which is more idiomatic. 65.26.203.135 (talk) 19:13, 29 August 2018 (UTC)

✅ Ok. —DIYeditor (talk) 20:16, 29 August 2018 (UTC)

Two new sections --- please help if you can
I just added two topics to the bottom of the "risks" section: AI and AI. Please help me find additional citations for these important topics, and help me link them into these topics elsewhere in Wikipedia. Thanks. CharlesGillingham (talk) 21:17, 29 July 2018 (UTC)

Also, at some point we may want to "move down" some of the material into the satellite articles (which are mostly pretty weak) and tighten up the "philosophy, ethics, speculation" section into short summaries with only the essential points. For example, I think the paragraph I just added for AI is plenty of material for this topic. The others could probably be cut down to a similar size. CharlesGillingham (talk) 21:21, 29 July 2018 (UTC)
 * I have removed the section on AI assisted tyranny. diff The problem is that it basically assumes that AI is a threat to one form of government, but not others. It also refers to a specific individual as a tyrant. Whatever we may think of the man, we don't do that here. Further, it puts a dark meaning to his words that I doubt he meant to portray. Perhaps there is a neutral manner of framing this. O3000 (talk) 11:50, 13 August 2018 (UTC)
 * , I see you are not by any means a new editor, but please carefully review Wikipedia polices on WP:NPOV and WP:OR. —DIYeditor (talk) 16:42, 13 August 2018 (UTC)




 * The threat to privacy and the targeted distribution of misinformation is currently being discussed pretty much everywhere (since the US 2016 election and Brexit) and is certainly essential topic if the article is to be up-to-date. Frankly, I did not think it was likely to be challenged.


 * I did source the Putin quote, and that article contains some of the essential ideas of the paragraph, although I'm sure there is a more detailed source that is a bit more on point. Couldn't that be enough of a source for something this widely discussed?


 * On Putin: the text did not say he is a tyrant. He is discussing the power that AI can confer to the state, and that is the point of the paragraph. I suppose you could make the implication that he is tyrant because he wants to use AI to confer more power to the state and to further the interests of Russia. The paragraph is fine without the Putin quote -- he just said it particularly forcefully, which helped reinforce the key ideas of the paragraph.


 * Finally on the term "democracy". I'm under the impression that the respect for privacy and the dislike of propaganda are aspects of a democratic government. What are the other forms of government that respect these rights? Certainly not communism, islamism, fascism, absolute monarchy, etc. What type of government were you thinking of? (Democratic socialism or democratic islamism count as democracy here, of course.)


 * At any rate, I'll take another stab at this with some more on-point sources. I don't have a lot of time for Wikipedia any more, so in the mean time could I just restore the title with an expand-section notice? The only thing I'm interested in here is making sure the article stays comprehensive we move forward. CharlesGillingham (talk) 20:59, 13 August 2018 (UTC)


 * Hi Charles. Just my 2 cents.  IMO section headers with nothing in them are not useful/ not a good idea.  Regarding the idea in general and the first 3/4 of that removed edit, it's basically hypothesizing about a very specific (assisting tyranny) one of the zillions of very narrow ways that it could do harm.   Then you follow with a very generic unrelated statement by Putin.  "Unrelated" unless the juxtaposition is an implicit statement that Putin is a tyrant, and whether you intended it or not it appears to be that. Such would violate several Wikipedia policies. Also, IMO the previously expressed concerns about the democracy wording are valid.  Maybe a section about more general increasing governmental power over individuals might be good? North8000 (talk) 22:00, 13 August 2018 (UTC)


 * Mostly right. You make three points above that I think are wrong, so I will find a reliable source that makes those points specifically. (1) It's not "hypothetical", because statistical machine learning is being used now to target individuals with propaganda -- the same algorithms that drive advertising to users are also used to drive propaganda to users. And because intelligence agencies around the world are customers of advanced machine learning for surveillance (voice recognition, face recognition, on-line activity monitoring). We're discussing only the uses of AI that are happening now; the other "jillions" remain hypothetical. (2) Russia is using statistical machine learning now for both surveillance & targeting propaganda, which is why the Putin quote is not "generic" or "unrelated". That is the specific context of the quote (but I didn't say that in the paragraph, so you all took the wrong implication.) (3) Democracy is the only form of government that I know of that doesn't condone propaganda or wide scale surveillance. I still need to see some kind of argument that there is another form of government that would have misgivings about using AI for propaganda or surveillance. Is there some other form of government I overlooked?


 * Again, forgive me for thinking that these points were unlikely to be challenged. I stand corrected. CharlesGillingham (talk) 11:14, 15 August 2018 (UTC)
 * Assuming that you consider the US a democracy, it certainly condones propaganda. (See the many mentions of propaganda at Voice of America.) Of course it’s only propaganda when someone else does it. As for surveillance, the NSA goes back 100 years and its computers now examine cross border traffic. Various police departments have made use of facial recognition systems examining random citizens on the streets. Major US corporations are collecting masses of information about web use, phone use, and movements of citizens. There are reports that services like Alexa and Google Home listen to conversations when not in use. Credit card companies keep records of every purchase used for sales programs. Supermarkets have been tracking each item you purchase since the ‘80s. As I browse sites, ads appear listing items I just purchased at Jet, or items I simply looked at on Bloomingdale’s site. (Geez, I’m going to start a conspiracy site.) O3000 (talk) 12:00, 15 August 2018 (UTC)

BTW, ten years ago this article had a section about the use of AI in the service of tyranny, which cited Pamela McCorduck's Machines Who Think and Aleksandr Solzhenitsyn's In the First Circle. At that time, it was hypothetical. But McCorduck and Solzhenitsyn's predictions have come true. CharlesGillingham (talk) 11:34, 15 August 2018 (UTC)

Charles, I think we're both knowledgeable and in agreement about what's actually happening. But by "narrow" and "hypothetical" I meant specifically to support tyranny. And Democracy is not the only alternative to tyranny (e.g. a benevolent dictatorship) nor a full solution to the issues. Now that I look at it, most of the problems revolve around just the selection of that one word "tyranny" North8000 (talk) 12:07, 15 August 2018 (UTC)


 * Well, we need a term that suggests harm, because this is a section on the risks associated with AI. Something like Abuse of AI by powerful bad actors, but not such a mouthful. Another issue: we're having -- by "democracy", I didn't mean the United States. Obviously the United States and US corporations are often bad actors as well, and often attempt to subvert the ideals of democracy. CharlesGillingham (talk) 20:18, 24 August 2018 (UTC)
 * Most any tool can be abused. The main difference with tech, and probably moreso AI, is the added flexibility, and the concomitant added opportunity for abuse. I think much of this is already covered in the article in general terms. Extension to potential abuses by government is obvious, and possibly no more a danger than misuse by other possible bad actors. O3000 (talk) 20:31, 24 August 2018 (UTC)
 * Maybe the area not fully covered is that AI is massively increasing the power of surveillance, and how it does so. Maybe with examples.  That might enhance the article in this area and solve everything without trying to get into characterizing misuses.  ? North8000 (talk) 21:58, 24 August 2018 (UTC)

What you're saying is true, but you're missing the main point: I think we have to cover the harm that AI is permitting today. CharlesGillingham (talk) 09:21, 30 August 2018 (UTC)
 * While your motivations are good here, I think we are faced with a degree of activism and a desire for how the article ought to read and what is important to tell the reader. The question is how reliable sources are addressing this issue. The way you want to phrase things is further indicative of having an agenda on your part. "Bad actors" and "tyranny" are highly subjective, and your association of propaganda and surveillance with being tyrannical is loaded. —DIYeditor (talk) 18:05, 30 August 2018 (UTC)