- Elon Musk has long criticized OpenAI and Sam Altman for keeping their technology closed source.
- However, Altman seemed to agree with Musk's views.
- Altman said Thursday that open-source AI models would strengthen America's advantage in the field.
OpenAI CEO Sam Altman has a vision for open source AI that's not far from what the company's former co-founder Elon Musk wanted.
Writing in The Washington Post on Thursday, Altman outlined a four-point plan to strengthen U.S. AI dominance by implementing safeguards, building infrastructure, restricting AI-related exports and forming a global community.
But what's notable about Altman's proposal is that it highlights the role that open-source AI models could play in strengthening American influence in the field.
“Making the open source model more readily available to developers in these countries will further strengthen our advantage,” Altman wrote. “The challenge of who will lead in AI is not just about exporting the technology, but about exporting the value that the technology supports.”
Musk has repeatedly criticized Altman's leadership since leaving the ChatGPT development company's board of directors in 2018. Musk has slammed Altman based on his position that OpenAI's systems are not as open as they should ideally be.
“OpenAI was founded as an open source (hence why I name it 'Open' AI) non-profit company to serve as a counterweight to Google, but is now effectively controlled by Microsoft as a closed source, profit-maximizing company,” Musk said in a February 2023 post on X.
“It was totally unintentional,” he added.
Musk filed a lawsuit against OpenAI in February, alleging that the company was violating its non-profit mission by partnering with Microsoft, but later dropped the lawsuit in June.
To be sure, Musk also has concerns about how open AI models should be: In private communications, the Tesla and xAI head seemed to agree with OpenAI's then-chief scientist Ilya Sutskever that there should be limits to openness.
“As we get closer to building AI, it makes sense that we will become less open,” Sutskever wrote to Musk in an email in January 2016.
“The openness of OpenAI means that once an AI is built, everyone should benefit from its results, but it's perfectly fine not to share the science (even though sharing everything is absolutely the right strategy in the short and probably medium term for adoption purposes),” Sutskever continued.
“Yes,” Musk replied.
Representatives for Altman and Musk did not immediately respond to Business Insider's request for comment sent outside normal business hours.
Altman's position on open source AI is more of an apology to Musk than anything else, as it's a recognition that that may be the direction the industry is heading.
On Tuesday, Meta CEO Mark Zuckerberg announced the company's latest AI model, Llama 3.1.
The model, which Meta says outperforms OpenAI's GPT-4o, is open source and available for free use to the public.
“We believe that open source is necessary to ensure a bright future for AI,” Zuckerberg said in a blog post on the same day.
“While debate continues about the safety of open source AI models, my view is that open source AI is safer than alternatives. Because open source makes the world richer and safer, governments will conclude that it is in their interest to support open source,” Zuckerberg wrote.
