Welcome back to The Prompt,
Elon Musk has his eyes on OpenAI. On Monday, Musk and a group of investors made a bid to buy OpenAI’s nonprofit arm for $97.4 billion and eventually merge it with his own AI company, xAI.
The offer comes as CEO Sam Altman tries to restructure OpenAI to become a for-profit benefit corporation and remove the non-profit board’s control over the company so that it can raise more capital. While it’s unclear whether Altman really has a say, he was quick to turn down Musk’s offer with a quip: “No thank you but we will buy twitter for $9.74 billion if you want,” he said on X. “OpenAI is not for sale,” he told Bloomberg, adding that the bid is Musk’s latest tactic to slow down the ChatGPT maker’s progress. The clash marks the latest turn in a longstanding feud between the two billionaires, which began when Musk left OpenAI in 2018 over a slew of disagreements about the company’s future.
Musk’s nearly $100 billion offer has resurfaced an important outstanding question: How much is OpenAI’s nonprofit arm worth or rather, how much should OpenAI board members’ stake in the company be worth?
Now let’s get into the headlines.
POLITICS
Speaking at the AI Action Summit in Paris, Vice President J.D. Vance urged world leaders to put fewer regulations on AI applications, adding that excessive regulation would “paralyze” the powerful technology. America’s leadership in AI headlined the speech and he said AI tools should be free from “ideological bias” and should not be used as a weapon for “authoritarian censorship.” So it’s not a complete surprise that the United States is not a signatory to a declaration to build inclusive, ethical, safe and sustainable artificial intelligence, which was signed by 60 other countries.
SHOW ME THE MONEY
Data unicorn Dbt Labs, whose systems helps companies clean and reformat their data for AI training, has crossed $100 million in annual recurring revenue (how much revenue a company expects to earn from current subscriptions.) CEO Tristan Handy told me that AI has been a “tailwind” for the company’s growth.
ETHICS + LAW
Thomson Reuters won a copyright infringement case it filed against legal AI startup Ross Intelligence in 2020, in which it claimed that the AI tool regurgitated the legal research firm’s material, Wired reported. That’s significant, considering that several AI startups facing similar lawsuits are using a similar “fair use” argument to defend their training practices.
AI DEAL OF THE WEEK
OpenAI cofounder Ilya Sutskear’s startup Safe Superintelligence is in talks to raise funding at a $20 billion valuation, according to Reuters. The company’s mission is to train safe AI models.
Also notable: Abu Dhabi-based AppliedAI has raised $55 million in funding, one of the largest Series A checks for the region’s startups. The company uses AI to automate back office tasks like processing insurance and healthcare claims.
DEEP DIVE
Thousands of artists are calling for London-based fine art auction house Christie’s to cancel an upcoming auction dubbed “Augmented Intelligence,” during which it plans to sell artworks created with AI tools like Dall-E, Midjourney and Stable Diffusion. Artists have expressed concerns that these AI art generators have been trained by scraping millions of copyrighted works of human artists from the internet, without their consent or giving them credit and compensation. Artwork produced with the help of these AI tools then compete with them in the market, affecting their livelihoods.
An open letter signed by over 4000 artists, urges Christie’s to call off its first AI auction, adding that it “further incentivizes AI companies’ mass theft of human artists’ work.” Karla Ortiz, an illustrator who’s part of a class action lawsuit against Runway ML, Stability AI and Midjourney for allegedly infringing copyright while training their photo and video-generating models, said artworks created entirely with AI cannot be copyrighted. The matter is “still in legal question” as the lawsuit progresses through court and the auction could be legally risky for the artists and the buyers, she said. “It’s just a kick in the face. It’s been upsetting,” Ortiz told me. “They [Christie’s] should’ve known better.”
In response to a request for comment, Christie’s spokesperson Jessica Stanley said, “The artists represented in this sale all have strong, existing multidisciplinary art practices, some recognized in leading museum collections. The works in this auction are using artificial intelligence to enhance their bodies of work.”
The auction will also feature work from artists that plan to combine AI art with other components such as a live performance. OpenAI’s first artist-in-residence Alexander Reben’s showcase will include a robot, powered by an AI model, that will, in real time, paint more of a canvas as the bids increase.
Leaders in the space point out that the auction is not a good look in the larger conversation around AI. Ed Newton Rex, CEO of Fairly Trained, a nonprofit which certifies AI companies for using fair training data, said selling AI artworks for tens of thousands of dollars would condone using AI tools that are trained on copyrighted works. “I think it’s a bit kind of grotesque, like having an auction to sell these things when the creators are saying, don’t use my work this way,” he said.
MODEL BEHAVIOR
OpenAI spent $14 million on its first ever advertisement which ran during the Super Bowl on Sunday. The 60 second ad, which showed black dots coalescing to create different scenes, drew parallels between ChatGPT and some of mankind’s greatest inventions such as the creation of the light bulb. But for some viewers, the commercial didn’t quite hit the mark. Instead, who found it to be too abstract. Ironically, the ad itself wasn’t generated by AI.
