This week in AI: OpenAI and publishers are partners of convenience

It's a tall order to keep up with an industry that's evolving as quickly as AI. So until an AI can do it for you, here's a handy overview of recent stories in the world of machine learning, along with notable research and experiments that we haven't covered on their own.

By the way, JS plans to launch an AI newsletter soon. Stay tuned. In the meantime, we're increasing the cadence of our semi-regular AI column, which used to be twice a month (or so), to weekly – so look out for more editions.

This week in AI, OpenAI announced that it has reached a deal with News Corp, the new publishing giant, to train generative AI models developed by OpenAI on articles from News Corp brands, including The Wall Street Journal, Financial times and MarketWatch. The agreement, which the companies describe as “multi-year” and “historic,” also gives OpenAI the right to display News Corp mastheads in apps like ChatGPT in response to certain questions – presumably in cases where the answers are partially or entirely sourced . from News Corp publications.

Sounds like a win for both sides, right? News Corp gets a cash injection for its content – ​​more than $250 million, Reportedly – at a time when the prospects of the media industry are even grimmer than usual. (Generative AI didn't help mattersmenacing significantly reduce publication referral traffic.) Meanwhile, OpenAI, which is battling copyright holders over fair use disputes on a number of fronts, has one less costly lawsuit to worry about.

But the devil is in the details. Keep in mind that the News Corp deal has an end date, as do all of OpenAI's content licensing deals.

That in itself is not bad faith on OpenAI's part. Granting perpetual licenses is a rarity in the media, given the motivation of all parties involved to keep the door open for renegotiation of the deal. However, it is a bit suspicious in light of OpenAI CEO Sam Altman's recent comments about the declining importance of AI model training data.

In an appearance on the “All-In” podcast, Altman said that he “definitely [doesn’t] I think there will be an arms race [training] data” because “when models get smart enough, at some point they shouldn't need more data anymore – at least not for training.” Elsewhere, he told MIT Technology Review's James O'Donnell says he's “optimistic” that OpenAI – and/or the broader AI industry – will “find a way out” [needing] more and more training data.”

Models are not yet as 'smart' as OpenAI leads to reportedly experimenting with synthetic training data and scour the far reaches of the web – and YouTube — for organic sources. But let's assume they will happen someday not A lot of additional data is needed to improve by leaps and bounds. Where does that leave publishers, especially after OpenAI scraped their entire archives?

The point I'm trying to make is that publishers – and the other content owners OpenAI has partnered with – seem to be partners of convenience in the short term, and little more. Through licensing agreements, OpenAI effectively neutralizes a legal threat – at least until the courts determine how fair use applies in the context of AI training – and gets to celebrate a PR victory. Publishers receive much-needed capital. And work on AI that could seriously harm these publishers continues.

Here are some other notable AI stories from the past few days:

• Spotify's AI DJ: Spotify's addition of its AI DJ feature, which introduces personalized song selections to users, was the company's first step toward an AI future. Now Spotify is developing an alternative version of that DJ who speaks Spanish, Sarah writes.
• Meta's AI Council: Meta announced on Wednesday the creation of a AI Advisory Board. However, there is a big problem: it only features white men. That feels a bit tone-deaf, since marginalized groups are the ones most likely to suffer the consequences of AI technology's shortcomings.
• FCC proposes AI disclosures: The Federal Communications Commission (FCC) has mandated that AI-generated content be disclosed in political advertising, but not banned. Devin has the full story.
• Respond to calls with your voice: Truecaller, the well-known caller ID service, will soon allow customers to use its AI-powered assistant to answer calls in their phone. own voice, thanks to a newly inked partnership with Microsoft.
• Humane is considering a sale: Humane, the company behind the much-hyped Ai Pin that launched last month to less than favorable reviews, is on the hunt for a buyer. The company has reportedly priced itself between $750 million and $1 billion, and the sale process is in the early stages.
• TikTok turns to generative AI: TikTok is the latest tech company to integrate generative AI into its advertising business, as the company announced Tuesday that it is launching a new TikTok Symphony AI suite for brands. The tools will help marketers write scripts, produce videos and improve their current advertising assets, Aisha reports.
• AI Summit in Seoul: At an AI security summit in Seoul, South Korea, government officials and AI industry executives agreed to implement basic security measures in the rapidly changing field and establish an international security research network.
• Microsoft's AI PCs: During a pair of keynotes at the annual Build developer conference this week, Microsoft unveiled a new line of Windows machines (and Surface laptops) it's calling Copilot+ PCs, plus generative AI-powered features like Recall, which will let users find apps, files and more . content they have viewed in the past.
• OpenAI's voting debacle: OpenAI is removing one of the voices in ChatGPT's text-to-speech feature. Users found the voice, named Sky, to be eerily similar to Scarlett Johansson (who has previously played AI characters) – and Johansson herself released a statement saying she had hired legal counsel to inquire about the Sky voice and exact to get details on how it was developed.
• British autonomous driving law: Britain's autonomous car regulations are now official after receiving royal assent, the final stamp that any legislation must pass before it becomes law.

More machine learning

A few interesting bits of AI-adjacent research for you this week. Prolific University of Washington researcher Shyan Gollakota strikes again with noise-canceling headphones you can ask about block everything but the person you want to listen to. While wearing the headphones, you press a button while looking at the person, and a sample is taken of the voice coming from that specific direction. This powers an auditory exclusion engine to filter out background noise and other voices.