- OpenAI has taken some heat for not being as “open” as its name might suggest.
- CEO Sam Altman explained the company’s shift toward closed AI models in a Reddit AMA on Thursday.
- It provides “an easier way to reach the security threshold,” and OpenAI wants to “open source more things in the future,” he said.
Why doesn’t OpenAI open source the AI models behind ChatGPT if the company is called OpenAI?
It comes down to what OpenAI is “good at” and the way to reach the company’s security limits for its AI models, CEO Sam Altman said in a Reddit Ask Me Anything session on Thursday.
“OpenAI has moved from a more open source approach to a more closed model in recent years,” said one Reddit user. “Can you elaborate on the reasoning behind this change and how you weigh the tradeoffs between openness and potential risks associated with widely accessible advanced AI technologies?”
Altman responded that open source “plays an important role in the ecosystem and there are great open source models in the world.”
“We also think there’s an important role in the world for robust and easy-to-use APIs and services, and given what we’re good at, we see an easier way to reach the safety threshold we want to achieve in this way,” he said. “We’re quite proud of the great value people get from our services.”
Altman added that he “would like us to open source more things in the future.”
The comments from Altman follow a year in which OpenAI has received criticism for not providing its AI models more openly. The company previously released code for its GPT-2 AI model in November 2019. However, it did not release code for its upcoming models, including its latest GPT-4o and o1-preview models.
Elon Musk, who co-founded OpenAI in 2015 with Altman and others before stepping down from the company’s board, has been one of the most vocal critics of its new approach.
In December 2022, a month after ChatGPT came out, he tweeted: “OpenAI started as open source and non-profit. Neither is true yet.” OpenAI became “limited profit”, a hybrid of non-profit and for-profit structures, in 2019.
Musk filed a lawsuit against Altman and OpenAI in March, alleging a breach of contract and that the company has abandoned its original mission. OpenAI published a blog post responding to his claims soon after.
An email exchange published in the blog post showed Ilya Sutskever, one of OpenAI’s co-founders and its former chief scientist, writing to Musk: “As we get closer to building AI, it will make sense to start being less open. Open in openAI means that everyone should benefit from the fruits of AI after it’s built, but it’s perfectly fine not to share the science (although sharing everything is definitely the right strategy in the short and possibly medium term for recruitment purposes).
The footage released by OpenAI showed Musk simply replying, “Yes.”
Musk dropped the lawsuit in June before filing a new lawsuit in August alleging that he was “duped” into co-founding the AI firm.
Being open source or closed source is a hot topic in the AI industry right now. One OpenAI competitor promoting its open source efforts is Meta, whose latest offering in the space, Llama 3.2 models, came out in September. CEO Mark Zuckerberg has said he believes open source is “safer than the alternatives” and “necessary for a positive AI future.”
It’s worth noting that the company requires firms with a user base of 700 million monthly active users or more to request the use of Llama, and Meta can decide whether to grant access or not.
The non-profit Open Source Initiative has previously said that this license “would not qualify as an open source license because it contains restrictions for commercial needs”. OSI also recently said that Meta’s Llama does not meet its definition of open source AI.
A Meta spokesperson said the company does not agree with OSI’s definition.
“There is no single definition of open source artificial intelligence, and defining it is a challenge because previous open source definitions do not capture the complexity of today’s rapidly advancing AI models,” the spokesperson said.
“We make Llama free and openly available, and our license and Acceptable Use Policy help keep people safe by imposing some restrictions,” they added. “We will continue to work with OSI and other industry groups to make AI more accessible and free responsibly, regardless of technical definitions.”
Elsewhere in the Reddit AMA, Altman said that OpenAI’s GPT-5 model likely won’t launch this year, and that he was using ChatGPT to answer some of the questions online.