Draft:Blue Jeans and Bloody Tears

Blue Jeans and Bloody Tears
Blue Jeans and Bloody Tears is a song composed and written by artificial intelligence (AI) as a homage to the Eurovision Song Contest. The song is performed by the esteemed singer Izhar Cohen, the winner of the Eurovision Song Contest 1978. It was released a week before the 64th edition of the Eurovision Song Contest.

Background
Motivated by a keen interest in exploring the intersection of artificial intelligence, music, and popular culture, the song pays homage to the diverse range of entries in the competition and pushes the boundaries of traditional musical composition, paving the way for innovative expressions within the Eurovision landscape.

The song was conceived and cretive directed by new media artist Nim Shapira. Shapira teamed up with Allenby Concept House to co-produce the project. Initially, Shapira wanted the song to represent Israel in that year's Eurovision Song Contest and approached Kan_11, the Israeli Public Broadcasting Corporation. However, the strict timeline for submitting a song in Israel (less than a few weeks) and the production of the song (several months) didn't allow for that. Shapira assembled a team of artists (himself and SulkyBunny), musicians (Avshalom Ariel, Amir Shoenfeld, Ben Scheflan), and programmers (Eran Hadas, Oracle Corporation). The team decided to collaborate under the pseudonym 'Sweaty Machines'.

It is noteworthy that this innovative undertaking transpired long before the term "AI" became commonly known, illustrating the project's visionary and pioneering approach in blending music with technology.

Lyrics and Melody
The song was curated and amalgamated by composer and songwriter Amir Shoenfeld. Additional team members were Ben Scheflan, Nim Shapira, and Eran Hadas. Together, they selected the most compelling elements to shape the final composition. Subsequently, a meticulous process of lyric selection ensued, where the team carefully chose lines that harmonized with the melody and adhered to the appropriate syllable count.

The song unfolds in a captivating paradox. . The verses delve into introspective themes of lost love and heartbreak through ballad-like singing by Izhar Cohen, with lines like "Baby, summer's gone, it's gone for the both of us." In stark contrast, the choruses erupt into vibrant electro beats, featuring high-pitched, computerized vocals chanting the paradoxical refrain "Blue jeans and bloody tears, there's no life without your life in misery." This push-pull tension between balladry and bangers reflects the song's self-aware and humorous nature.

The lyrics, meticulously chosen by the team, further emphasize this ironic stance. While grammatically correct, they embrace deliberate weirdness and occasional lack of literal meaning. Phrases like "In this pain of the bird who's flying over roses of sadness" prioritize emotional resonance and bizarre charm over conventional narrative logic. This approach playfully acknowledges the song's AI-generated origin, reveling in the "uncanny valley" effect evoked by the lyrics' unconventional nature. Additional notable lines include: "tears will always have wet eyes" and "there's no life without your life in misery."

Musical Production
The song's production was overseen by musician Avshalom Ariel, a notable contributor to the Eurovision Song Contest 2018 winning song Toy_(song).

Adding to this unique musical tapestry is the song's structure as a duet. Cohen's emotive tenor anchors the introspective verses, while the high-pitched, computerized vocals in the choruses enhance the sense of uncanny valley and AI authorship. This juxtaposition between the familiar warmth of Cohen's voice and the robotic precision of the synthetic vocals adds another layer of complexity to the song's exploration of love, loss, and the ever-blurring lines between human and machine.

In a manner reminiscent of the Eurovision Song Contest's characteristic charm and kitsch, the song was infused with two musical modulations towards its conclusion. This stylistic choice, akin to the conventions observed in Eurovision songs, reflects a deliberate nod to the competition's distinctive musical aesthetics, enhancing the overall composition with a touch of familiarity and thematic resonance.

Lyrics Generation
Led by poet and software developer Eran Hadas, the lyric generation process underwent meticulous stages. Initially, data preparation involved the thorough cleaning and filtering of hundreds of Eurovision songs, eliminating place names, person names, repetitions, special characters, and questions from the dataset. Subsequently, model training ensued, evaluating various machine learning models, predominantly focusing on LSTM Recurrent Neural Networks. This stage included the fine-tuning of models through experimentation with different hyperparameters to optimize their performance. The final phase involved line generation, wherein thousands of lines of text were created, with a careful emphasis on ensuring none were identical to any of the input songs, thereby guaranteeing the originality of the lyrical content.

Melody Generation
The intricate process overseeing the musical composition of "Blue Jeans and Bloody Tears," orchestrated by Oracle Corporation in Israel and the UK, entailed the generation of a multitude of melody tunes through the application of advanced artificial intelligence algorithms. A discerning approach to melody selection was meticulously executed, influenced by song statistics and exhaustive research on the historical performance of Eurovision entries.

During the production phase, a deliberate emphasis on Eurovision-style elements characterized the meticulous efforts to ensure the final song resonated authentically within the established norms of the song competition. Notably, the team manually inserted hundreds of songs into a comprehensive database, detailing key, major/minor tonality, BPM, genre, and other pertinent attributes. Each song was further enriched with MIDI files, systematically constructed to include introductory sections, verses, choruses, and c-parts.

For the melody creation, a collaborative effort with ORACLE was undertaken, involving the development of custom Python code employing two open-source machine learning libraries—LSTM and GAN. This intricate process was facilitated on GPU and HPC infrastructure on the Oracle Cloud platform.

Music Video
The music video was crafted by the esteemed BAFTA & CANNES Palm d'Or-nominated UK-based animation studio SulkyBunny, led by Karni Arieli and Saul Freed.

The music video for "Blue Jeans and Bloody Tears" showcases an animated mandrill monkey-shaped robot whose belly prints out either text or calculations within a limbo-like environment. The video concludes with a display of fireworks, adding lively and visually captivating elements to the unique and memorable nature of the video.

Reception and Impact
While "Blue Jeans and Bloody Tears" did not officially compete in the Eurovision Song Contest, it swiftly gained cult status, attracting significant media attention    and sparking discussions about the role of AI in art and music. This project demonstrated the potential of AI as a creative tool, paving the way for new avenues in song composition and artistic expression.

The song's release marked an unprecedented success, rapidly capturing the attention of global audiences. The video, infused with humor and catchiness, achieved millions of views on YouTube within hours, all without the aid of any paid promotion.

Awards and memorable mentions
The project was showcased at the 2023 TED conference. It preceded TikTok CEO Shou Zi Chew talk with head of TED Chris Anderson.

The song was discussed in "A Song for Europe: Popular Music and Politics in the Eurovision Song Contest" by Ivan Raykoff and Robert Deam Tobin.

The song received several awards such as the 2020 Webby Award Honoree Art & Experimental, the 2019 Lovie Awards Winner (The European “Webby” Awards): Gold in Internet Video: Weird & Experimental; Winner: Bronze in Internet Video: Viral Video.

It was also part of the 2019 The Wrong Biennale and the 2019 NeurIPS Creativity Workshop.

These awards underscored the song's ability to transcend conventional boundaries, establishing "Blue Jeans and Bloody Tears" as a pioneering force in the evolving landscape of AI-generated art and music. The project's success not only resonated with music enthusiasts but also contributed significantly to the broader conversation about the role of AI in shaping the future of artistic endeavors.

Major media outlets, including Bloomberg, covered the phenomenon. The song even made its way into the Swedish late-night show Svenska Nyheter. Fans created fan art and unofficially incorporated it into VR games like Synth Riders and Beat Saber.

Legacy
A year after the song's release, VPRO the Netherlands' Public Broadcasting Channel initiated a User-Generated Content (UGC) competition, inviting fans worldwide to create AI-generated songs as an homage to the innovative competition

In late 2023, The Eurovision Song Contest considered banning artificial intelligence (AI) from participating in the event due to concerns about its impact on creativity in music. Jean-Philip de Tender, from the European Broadcasting Union, highlighted the need for human creativity in songwriting and performance. The music industry has seen debates around AI-generated content, with Universal Music Group and YouTube collaborating to establish guidelines.