This story was initially published as part of tech News Now, TheStreet's daily tech digest.
In 2015, Elon Musk, fearful of the existential threat of artificial intelligence, teamed up with Sam Altman and a few other Silicon Valley people to launch a nonprofit ai research lab called OpenAI.
The lab's mission, focused on openness and transparency, was to develop artificial general intelligence (AGI) to benefit humanity, rather than maximize shareholder profits. It was designed to be a transparent counterweight to Google. Since that noble beginning, Musk and his millions have left the company, which is now led by Altman, and OpenAI's bylaws appear to have changed somewhat.
The company is now a hybrid nonprofit and for-profit limited combination. And in Musk's absence, OpenAI (in addition to marketing its products, obscuring transparency efforts, and shutting down its technology) turned to Microsoft for funding, receiving a $13 billion investment.
Related: Tesla boss Musk is pleased with OpenAI's 'Game of Thrones' outcome
in a lawsuit filed on February 29, Musk has accused OpenAI of breach of contract, breach of fiduciary duties, and unfair business practices, among other things. He seeks a legal requirement that forces OpenAI to return to its original founding charter, namely transparency and making ai research available to the public.
Musk is also seeking restitution of all money he invested in OpenAI while they engaged in the “unfair” practices described in the lawsuit, as well as general, compensatory and punitive damages.
Deeper ai Analysis:
- The director of the think tank warns of the danger of “non-democratic technological leaders deciding the future”
- George Carlin resurrected, without permission, by the self-proclaimed 'Comedy ai'
- artificial intelligence is a sustainability nightmare, but it doesn't have to be
“OpenAI has become a de facto closed-source subsidiary of the world's largest technology company: Microsoft,” the lawsuit states. “Under his new Board, he is not only developing but actually refining AGI to maximize Microsoft's profits, rather than for the benefit of humanity.”
ai researcher Gary Marcus said in a post which agrees with Musk's general description of the facts: “By any reasonable metric, OpenAI as a whole is no longer functioning according to the mission it originally set out to do. Elon did not get what he paid for and worked for.” .
“I won't predict who will win, but I will say that Elon is right. And anyone doing deals with Altman would do well to read him,” Marcus said. “The company Sam and Greg built has little to do with what was originally promised.”
The lawsuit also points to the boardroom drama that emerged from OpenAI last year, when Altman was fired and then reinstated over the course of a weekend. The lawsuit says that “upon information and belief, Altman's termination was due in part to OpenAI's advancement in realizing AGI,” and adds that Altman's reinstatement was the result of Microsoft's “coercive power” over OpenAI. and its board of directors.
“With the reinstatement of Mr. Altman and the restructuring of the Board, OpenAI's corporate structure, which had been designed as a system of checks and balances between the nonprofit arm, the for-profit arm, the Board and CEO to ensure a lucrative mission was being carried out, collapsed overnight,” the suit says.
At the time, Musk expressed intense concern about the boardroom changes taking place at his former startup.
OpenAI did not respond to a request for comment.
Related: ai Tax Fraud: Why It's So Dangerous and How to Protect Yourself From It
The AGI of everything
The lawsuit further argues that OpenAI's GPT-4, which powers its main version of ChatGPT, is “capable of reasoning,” something researchers disagree with (and something that is difficult to investigate, given the lack of OpenAI transparency).
An article published online by ai researcher Melanie Mitchell in early February found that “GPT models still lack the kind of abstract reasoning necessary for human-like fluid intelligence.”
A separate document published in August, which has not yet been peer-reviewed, argues that GPT-4 “cannot reason.”
frame referred to GPT-4 as “one more giant step for hype, but not necessarily one giant step for science, AGI, or humanity.”
“There are no scientific publications describing the design of GPT-4,” the lawsuit says, echoing Marcus's earlier writings. “Instead, there are only press releases boasting performance. In terms of information and belief, this secrecy is mainly due to commercial considerations, not security.”
The complaint says Musk contributed a total of $44 million to OpenAI between 2016 and 2020. The company is now valued at around $86 billion.
“But while some, like Mr. Musk, see an existential threat in AGI, others see AGI as a source of profit and power,” the lawsuit says.
Related: Experts explain problems with Elon Musk's ai security plan
Musk, ai and business
Meanwhile, Musk is doing a lot of his own work in artificial intelligence, from self-driving cars to Tesla's Optimus robot and Grok, a large language model designed to compete with ChatGPT. In fact, many analysts and investors see ai as an integral part of Tesla's business.
Six months after signing a letter calling for a pause in the development of advanced ai systems, Musk launched xAI, which Grok later launched, saying at the time that his vision for safe ai involves “growing” curious ai and that Find the truth”. model.
Two experts TheStreet spoke to at the time noted their skepticism about this approach.
“I don't think attributing human attributes to ai models is a good idea, nor accurate in any way. Models cannot be curious because they are not sentient,” ai expert and researcher. Dr. Sasha Luccioni he told The Street in July.
Musk has also hinted that Tesla, through its autonomous driving efforts, has “figured out some aspects of AGI.” saying in August: “The car has a mind. Not a huge mind, but a mind nonetheless.”
Many researchers, however, have questioned whether AGI is possible at all, with some claiming that such hype around AGI's dangerous or benevolent world-changing potential is little more than a ploy for power.
Contact Ian with ai tips and stories by email, [email protected] or Signal at 732-804-1223.
Related: Meet Sam Altman, the man behind OpenAI's revolutionary ChatGPT