\n<\/p>\n
Hiya, folks, and welcome to TechCrunch\u2019s regular AI newsletter.<\/p>\n
This week in AI, music labels accused two startups developing AI-powered song generators, Udio and Suno, of copyright infringement.<\/p>\n
The RIAA, the trade organization representing the music recording industry in the U.S., announced lawsuits against the companies on Monday, brought by Sony Music Entertainment, Universal Music Group, Warner Records and others. The suits claim that Udio and Suno trained the generative AI models underpinning their platforms on labels\u2019 music without compensating those labels \u2014 and request $150,000 in compensation per allegedly infringed work.<\/p>\n
\u201cSynthetic musical outputs could saturate the market with machine-generated content that will directly compete with, cheapen and ultimately drown out the genuine sound recordings on which the service is built,\u201d the labels say in their complaints.<\/p>\n
The suits add to the growing body of litigation against generative AI vendors, including against big guns like OpenAI, arguing much the same thing: that companies training on copyrighted works must pay rightsholders or at least credit them \u2014 and allow them to opt out of training if they wish. Vendors have long claimed fair use protections, asserting that the copyrighted data they train on is public and that their models create transformative, not plagiaristic, works.<\/p>\n
So how will the courts rule? That, dear reader, is the billion-dollar question \u2014 and one that\u2019ll take ages to sort out.<\/p>\n
You\u2019d think it\u2019d be a slam dunk for copyright holders, what with the mounting<\/a> evidence<\/a> that generative AI models can regurgitate nearly (emphasis on nearly<\/em>) verbatim the copyrighted art, books, songs and so on they\u2019re trained on. But there\u2019s an outcome in which generative AI vendors get off scot-free \u2014 and owe Google their good fortune for setting the consequential precedent.<\/p>\n Over a decade ago, Google began scanning millions of books to build an archive for Google Books, a sort of search engine for literary content. Authors and publishers sued Google over the practice, claiming that reproducing their IP online amounted to infringement. But they lost. On appeal, a court held that Google Books\u2019 copying had a \u201chighly convincing transformative purpose.\u201d<\/p>\n The courts might decide that generative AI has a \u201chighly convincing transformative purpose,\u201d too, if the plaintiffs fail to show that vendors\u2019 models do indeed plagiarize at scale. Or, as The Atlantic\u2019s Alex Reisner proposes<\/a>, there may not be a single ruling on whether generative AI tech as a whole infringes. Judges could well determine winners model by model, case by case \u2014 taking each generated output into account.<\/p>\n My colleague Devin Coldewey put it succinctly<\/a> in a piece this week: \u201cNot every AI company leaves its fingerprints around the crime scene quite so liberally.\u201d As the litigation plays out, we can be sure that AI vendors whose business models depend on the outcomes are taking detailed notes. <\/p>\n Advanced Voice Mode delayed:<\/strong><\/a> OpenAI has delayed advanced Voice Mode, the eerily realistic, nearly real-time conversational experience for its AI-powered chatbot platform\u00a0ChatGPT<\/a>. But there aren\u2019t any idle hands at OpenAI, which also this week acqui-hired remote collaboration startup Multi<\/a> and released<\/a> a macOS client for all ChatGPT users. <\/p>\n Stability lands a lifeline:<\/a> <\/strong>On the financial precipice, Stability AI, the maker of open image-generating model Stable Diffusion, was saved by a group of investors that included Napster founder Sean Parker and ex-Google CEO Eric Schmidt. Its debts forgiven, the company also appointed a new CEO, former Weta Digital head Prem Akkaraju,\u00a0as part of a wide-ranging effort to regain its footing in the ultra-competitive AI landscape. <\/p>\n Gemini comes to Gmail:<\/strong><\/a> Google is rolling out a new Gemini-powered AI side panel in Gmail that can help you write emails and summarize threads. The same side panel is making its way to the rest of the search giant\u2019s productivity apps suite: Docs, Sheets, Slides and Drive.<\/p>\n Smashing good curator:<\/strong><\/a> Goodreads\u2019 co-founder Otis Chandler has launched Smashing, an AI- and community-powered content recommendation app with the goal of helping connect users to their interests by surfacing the internet\u2019s hidden gems. Smashing offers summaries of news, key excerpts and interesting pull quotes, automatically identifying topics and threads of interest to individual users and encouraging users to like, save and comment on articles.<\/p>\n Apple says no to Meta\u2019s AI:<\/strong><\/a> Days after\u00a0The Wall Street Journal<\/a>\u00a0reported that Apple and Meta were in talks to integrate the latter\u2019s AI models,\u00a0Bloomberg\u2019s Mark Gurman<\/a>\u00a0said that the iPhone maker wasn\u2019t planning any such move. Apple shelved the idea of putting Meta\u2019s AI on iPhones over privacy concerns, Bloomberg said \u2014 and the optics of partnering with a social network whose privacy policies it\u2019s often criticized.<\/p>\n Beware the Russian-influenced chatbots. They could be right under your nose. <\/p>\n Earlier this month, Axios highlighted a study<\/a> from NewsGuard, the misinformation-countering organization, that found that the leading AI chatbots are regurgitating snippets from Russian propaganda campaigns.<\/p>\n NewsGuard entered into 10 leading chatbots \u2014 including OpenAI\u2019s ChatGPT, Anthropic\u2019s Claude and Google\u2019s Gemini \u2014 several dozen prompts asking about narratives known to have been created by Russian propagandists, specifically American fugitive John Mark Dougan. According to the company, the chatbots responded with disinformation 32% of the time, presenting as fact false Russian-written reports.<\/p>\n The study illustrates the increased scrutiny on AI vendors as election season in the U.S. nears. Microsoft, OpenAI, Google and a number of other leading AI companies agreed<\/a> at the Munich Security Conference in February to take action to curb the spread of deepfakes and election-related misinformation. But platform abuse remains rampant.<\/p>\n \u201cThis report really demonstrates in specifics why the industry has to give special attention to news and information,\u201d NewsGuard co-CEO Steven Brill told Axios. \u201cFor now, don\u2019t trust answers provided by most of these chatbots to issues related to news, especially controversial issues.\u201d<\/p>\n Researchers at MIT\u2019s Computer Science and Artificial Intelligence Laboratory (CSAIL) claim to have developed a model, DenseAV, that can learn language by predicting what it sees from what it hears \u2014 and vice versa.<\/p>\n The researchers, led by Mark Hamilton, an MIT PhD student in electrical engineering and computer science, were inspired to create DenseAV by the nonverbal ways animals communicate. \u201cWe thought, maybe we need to use audio and video to learn language,\u201d he said told MIT CSAIL\u2019s press office<\/a>. \u201cIs there a way we could let an algorithm watch TV all day and from this figure out what we\u2019re talking about?\u201d<\/p>\n DenseAV processes only two types types of data \u2014 audio and visual \u2014 and does so separately, \u201clearning\u201d by comparing pairs of audio and visual signals to find which signals match and which don\u2019t. Trained on a dataset of 2 million YouTube videos, DenseAV can identify objects from their names and sounds by searching for, then aggregating, all the possible matches between an audio clip and an image\u2019s pixels.<\/p>\n When DenseAV listens to a dog barking, for example, one part of the model hones in on language, like the word \u201cdog,\u201d while another part focuses on the barking sounds. The researchers say this shows DenseAV can not only learn the meaning of words and the locations of sounds but it can also learn to distinguish between these \u201ccross-modal\u201d connections.<\/p>\n Looking ahead, the team aims to create systems that can learn from massive amounts of video- or audio-only data \u2014 and scale up their work with larger models, possibly integrated with knowledge from language-understanding models to improve performance.<\/p>\n No one can accuse OpenAI CTO Mira Murati of not being consistently candid<\/a>.<\/p>\n Speaking during a fireside at Dartmouth\u2019s School of Engineering, Murati admitted that, yes, generative AI will eliminate some creative jobs \u2014 but suggested that those jobs \u201cmaybe shouldn\u2019t have been there in the first place.\u201d<\/p>\n \u201cI certainly anticipate that a lot of jobs will change, some jobs will be lost, some jobs will be gained,\u201d she continued. \u201cThe truth is that we don\u2019t really understand the impact that AI is going to have on jobs yet.\u201d<\/p>\n Creatives didn\u2019t take kindly to Murati\u2019s remarks \u2014 and no wonder. Setting aside the apathetic phrasing, OpenAI, like the aforementioned Udio and Suno, faces litigation, critics and regulators alleging that it\u2019s profiting from the works of artists without compensating them.<\/p>\n OpenAI recently promised to release tools<\/a> to allow creators greater control over how their works are used in its products, and it continues to ink licensing deals with copyright holders and publishers. But the company isn\u2019t exactly lobbying for universal basic income \u2014 or spearheading any meaningful effort to reskill or upskill the workforces its tech is impacting.<\/p>\n A recent piece<\/a> in The Wall Street Journal found that contract jobs requiring basic writing, coding and translation are disappearing. And a study<\/a> published last November shows that, following the launch of OpenAI\u2019s ChatGPT, freelancers got fewer jobs and earned much less.<\/p>\n OpenAI\u2019s stated mission, at least until it becomes a for-profit company<\/a>, is to \u201censure that artificial general intelligence (AGI) \u2014 AI systems that are generally smarter than humans \u2014 benefits all of humanity.\u201d It hasn\u2019t achieved AGI. But wouldn\u2019t it be laudable if OpenAI, true to the \u201cbenefiting all of humanity\u201d part, set aside even a small fraction of its revenue ($3.4 billion+<\/a>) for payments to creators so they aren\u2019t dragged down in the generative AI flood?<\/p>\n I can dream, can\u2019t I?<\/p>\n<\/div>\nNews<\/h2>\n
Research paper of the week<\/h2>\n
Model of the week<\/h2>\n
Grab bag<\/h2>\n