Elon Musk is suing OpenAI and co-founders Sam Altman and Greg Brockman for allegedly violating the contract they signed when founding OpenAI. raised.
This lawsuit provides an interesting look at the birth of a company that has become a leader in developing cutting-edge AI models. This hinges on OpenAI’s two main goals, which are documented in the company’s “founding agreement.”
The three agreed that OpenAI “will be a non-profit organization developing AGI for the benefit of humanity rather than a for-profit company seeking to maximize shareholder profits.”
They also agreed that OpenAI “will be open source, balancing only countervailing safety considerations and not keeping the technology private and secret for proprietary commercial reasons.”
Musk was and still is fearful that the profitable pursuit of AGI and monopoly control of technology by companies like Google could be detrimental to human society.
The idea behind the agreement was that OpenAI would create open and secure AGI that could be freely used by humanity, eliminating the risk that commercial companies would use the technology to harm humanity.
violation
The undisputed leading AI model today is OpenAI’s GPT-4, but despite the name of the company that developed it, nothing has been made public. Unlike GPT-3, OpenAI hasn’t even released a technical document detailing how GPT-4 works.
Although the founding agreement states that the company will pursue open source technology, only OpenAI, and possibly Microsoft, will have access to GPT-4’s code, weights, and training data.
It costs money to run a company like OpenAI, so it’s not surprising that they charge users to use their products. But an exclusive licensing deal with Microsoft means that model is now used to generate profits for Big Tech. It’s the exact opposite of the founding principles that Musk said they all agreed to.
“OpenAI, Inc. has effectively been transformed into a closed-source subsidiary of Microsoft, the world’s largest technology company,” the lawsuit states. “The new Board of Directors is actually improving, not just developing, AGI to maximize Microsoft’s interests, not the interests of humanity.”
Is there AGI?
Microsoft’s license agreement applies only to OpenAI’s pre-AGI technology. Once an OpenAI model is determined to have reached AGI level, the licensing agreement prohibits Microsoft from accessing it.
Musk’s lawsuit makes the bold claim that “GPT-4 is an AGI algorithm and therefore Microsoft should not have access to it.”
It also confirms rumors about what OpenAI’s engineers are working on. “OpenAI is currently developing a model called Q* (Q star) that has even stronger claims for AGI,” the lawsuit claims.
Ultimately, the OpenAI Board of Directors determines what qualifies as an AGI under the license agreement. Following Altman’s firing, rehiring and board replacement, the decision now rests with new board members appointed with Microsoft’s approval.
Is this why Ilya Sutskever was removed from the board and quietly locked in a back room? Has anyone heard from him in months? What did Ilya see?
Don’t expect the new board to rush to assert that AGI has been achieved, even though that is the company’s stated goal.
Fear of AI
One of the cited motivations behind the lawsuit is Musk’s ongoing fear of AGI falling into the wrong hands.
The lawsuit details how Musk met with DeepMind co-founder Demis Hassabis in 2012 to discuss these fears, some of which Hassabis also shared.
It also states, “After a meeting with Mr. Hassabis and DeepMind’s investors, one of the investors stated that the best thing they could do for humanity was to shoot Mr. Hassabis on the spot.”
When Musk discussed his fears with Google’s then-CEO Larry Page, Page said that superintelligent AI replacing humans was just “the next step in evolution,” and Musk responded with his pro-human views. He claimed, “Seed.”
Is Musk being overly dramatic, or should we be worried?
If you want to know the story behind how a big part of the industry got to where it is today, the lawsuit is worth reading. But it doesn’t look like this case will be solved anytime soon.
OpenAI may confirm Musk’s fears about AI and use this as his main reason for not releasing its models as open source. Asserting the company’s interests may be more challenging for OpenAI’s already busy legal team.
