As artificial intelligence remains the hottest topic of 2022, last month President Biden stood alongside big tech leaders as they pledged impotent “voluntary commitments” to control the emerging technology. Those leaders did so fully comprehending the dangers posed by the rapid and unrestricted penetration of AI through society — which are especially grave for artists and creators.
None of the commitments can give anyone in the music or creative industries comfort. They are as basic and malleable as “prioritizing research” and “sharing information on managing AI risks.” It is impossible to monitor big tech’s compliance with them. Even worse, all of the commitments are unenforceable.
We are already experiencing the consequences of the unbridled development and use of AI. Copyrighted material is being routinely ingested and used by AI conglomerates without the consent or even knowledge of rights holders. AI-generated vocals and deepfakes — given prominence by the Fake Drake and The Weeknd’s “Heart On My Sleeve” saga — are prevalent and becoming more realistic by the day. “Artificial streaming” — whereby AI bots create and upload songs, and then artificially inflate streaming numbers — is a massive issue for the streaming industry. Misinformation and inaccuracies in AI output are rampant. Generative AI programs suffer from what experts call “hallucination” — where they make up or misrepresent facts. The victims are widespread, ranging from lawyers given fake legal cases to cite in court papers, to a professor named as the accused in a sexual harassment scandal fabricated by AI.
The lacuna of AI regulation and long wait for decisions in significant AI court cases leaves rights holders, and broader society, at the mercy of big tech. That includes the protagonists of the 2016 Cambridge Analytica scandal (where there were laws in place prohibiting such misconduct), Frances Haugen’s 2021 revelations about Facebook and its platforms’ impact on issues such as teen health and even human trafficking, as well as those being investigated by the FTC for engaging in unfair or deceptive practices causing harm to consumers. Some in this group have recently shown immense hostility to lawmakers, and threatened to leave the EU if it regulates AI. None are those whose history or current actions compel public trust.
We also have minimal, if any — a point recently illustrated by the California Stability AI lawsuit, in which key issues are copyright infringement and data scraping — transparency around the ingestion and scraping of data by companies that own generative AI. Content and data protections were not formulated with current forms of AI in mind. Nor were copyright and right of publicity laws. It is unclear whether they provide sufficient protections, and in any case, are difficult to enforce amidst the black box of AI data ingestion.
Draconian regulation is not the answer to these issues. Nor is inaction. Congress has been exceptionally slow to move. While some lawmakers have proposed legislation, held a few congressional hearings, and suggested new federal agencies to deal with AI, nothing meaningful has resulted. The significant AI litigation is also not progressing quickly. Recently, a Federal Judge indicated that he was inclined to dismiss the California Stability AI lawsuit, but would give the plaintiffs a chance to reformulate their case. This means any decision or guidance is unlikely this year. Similar court cases will not provide definitive guidance soon, and many may settle. Even if decisions are released, they will not be universally applicable.
U.S. Congress’ inaction starkly contrasts the European Union’s continued development of its “AI Act,” which includes important guardrails for the use of AI. For generative AI such as ChatGPT, the AI Act requires disclosure of content generated by AI; prohibitions on the ability to generate illegal content; and publication of copyrighted data used for training. It also prohibits real-time and remote biometric identification systems and cognitive behavioral manipulation using AI. While the final form of the Act is being negotiated, the EU hopes to “reach an agreement by the end” of 2023.
However, the EU AI Act is not a panacea — especially not for artists and creators, as music industry organizations like GESAC, ICMP, IFPI, IMPALA and IMPF recently pointed out to the EU. To properly protect human creativity and rights in creative output, more can be done to increase transparency regarding the data on which AI is trained and record keeping of the same — particularly content which is protected by registered copyrights. Such publicly accessible information will enable artists and creators to determine if their content has been ingested by AI, and also to make fully informed assessments as to whether AI outputs constitute infringement or fair use of their content. Stronger protections around the use of video and audio deepfakes and digital recreations of humans, particularly celebrities, are becoming an increasing priority to protect privacy, creativity and livelihoods.
The stakes are too high for a wait and see approach. We saw what happened when that approach was employed with cryptocurrency: FTX and other similar debacles. The potential impact of generative AI is greater because of its universal application and almost limitless potential. As history and the cacophony of current AI lawsuits make clear, big tech has little regard for the intellectual property, livelihoods and creativity of artists and creators, nor for individuals’ privacy and personal information.
The recent “voluntary commitments” won’t change a thing. Congress should take swift action. It has a unique opportunity to lead the world in AI regulation by passing an enhanced version of the EU AI Act. The risks of inaction amidst the rapid development and use of generative AI — like AI’s capabilities — are in many respects, existentially threatening.
Nicholas Saady is a U.S. and Australian lawyer, who represents high-profile organizations and individuals — including major artists, labels and agents — regarding complex intellectual property, technology and commercial matters. He has also published widely on issues relating to technology, AI, NFTs and cryptocurrency.
