Sam Altman hopped onto X on Saturday night and told users to ask him anything about OpenAI’s agreement with the Pentagon.
Altman, late on Friday, announced that his company had finalized a deal with the Department of War to use its AI models. OpenAI’s deal came after Anthropic refused an ultimatum regarding the terms of use of its frontier model, Claude, for deployment in mass domestic surveillance and fully autonomous weapons.
Here are 5 big takeaways from Altman’s AMA.
The OpenAI-Pentagon deal was ‘rushed,’ and Altman knows the ‘optics’ don’t look good
The Pentagon deal was done quickly in “an attempt to de-escalate the situation,” Altman wrote on X.
He added in a separate post that the deal had been “rushed.”
Still, the “optics don’t look good” for OpenAI, he wrote.
“If we are right and this does lead to a de-escalation between the DoW and the industry, we will look like geniuses, and a company that took on a lot of pain to do things to help the industry,” he wrote.
“If not, we will continue to be characterized as rushed and uncareful,” he wrote.
Altman added that he sees “promising signs” for where this will all land for OpenAI.
OpenAI took the Pentagon deal because it ‘got comfortable’ with the ‘contract language’
Altman was asked why the Department of War went with OpenAI over Anthropic. He said he wouldn’t speak for his competitor, but did speculate on why OpenAI got the contract inked first.
“First, I saw reporting that they were extremely close on a deal, and for much of the time both sides really wanted to reach one,” Altman wrote. “I have seen what happens in tense negotiations when things get stressed and deteriorate super fast, and I could believe that was a large part of what happened here.”
He added that OpenAI and the Department of War “got comfortable with the contractual language” as well.
“I think Anthropic may have wanted more operational control than we did,” he added.
OpenAI has 3 redlines, but it’s open to changing them as tech evolves
Altman said that OpenAI has “three redlines.” But those redlines could change — and there could be more of them put in place — as the technology evolves, and “new risks” come into play.
“But a really important point: we are not elected. We have a democratic process where we do elect our leaders,” Altman wrote. “We have expertise with the technology and understand its limitations, but I think you should be terrified of a private company deciding on what is and isn’t ethical in the most important areas.”
“Seems fine for us to decide how ChatGPT should respond to a controversial question,” he added. “But I really don’t want us to decide what to do if a nuke is coming towards the US.”
Altman says Anthropic is on a ‘dangerous’ path
Altman said OpenAI had been talking to the Department of War for “many months” about non-classified work, before “things shifted into high gear on the classified side.”
“We found the DoW to be flexible on what we needed, and we want to support them in their very important mission,” he wrote.
“I think the current path things are on is dangerous for Anthropic, healthy competition, and the US,” Altman wrote on X as well. “We negotiated to make sure similar terms would be offered to all other AI labs.”
He also asked for “some empathy” for the Department of War, given its “extremely important mission.”
And, in Altman’s words:
Altman says AI can help counter big security threats on two fronts
Altman says AI could come in useful on two fronts. Firstly, the US’s “ability to defend against major cyber attacks,” particularly, an attack that might take down the country’s electrical grid.
Secondly, biosecurity is an area where AI could help.
“I do not think we are currently set up well enough to detect and respond to a novel pandemic threat,” Altman said.
Read the original article on Business Insider
The post 5 big takeaways from Sam Altman’s Saturday night AMA on OpenAI’s Pentagon deal appeared first on Business Insider.
