OpenAI CEO Sam Altman went into full damage control mode over the weekend. A day before the United States attacked Iran, the embattled C.E.O. announced The company has signed a new agreement with the Pentagon to limit how its AI models can be used — and the blow is clearly affecting the company’s bottom line, as Altman sounds deeply defensive.
Many users saw the military terms move as an attempt to snatch billions of dollars of government contracts from the clutches of its rival Anthropic. Last week, Anthropic CEO Dario Amodei refused to accede to the Defense Department’s demands, draw a line in the sand And it is insisting that its AI models cannot be used to create autonomous killing machines or mass surveillance of Americans, a decision that has been applauded by many users of its chatbot cloud.
Despite the genuineness of Amodei’s constant assurances – there are plenty of reasons not to take the billionaire CEO at his word – OpenAI effectively gave Anthropic a huge PR win. The changing dynamics triggered a mass exodus from OpenAI’s ecosystem, along with uninstall rates for OpenAI’s ChatGPT. 295 percent increase day after day on Saturday, the day after OpenAI announced its deal with the Pentagon.
Now, Altman is continuing his apology tour, admitted in a long tweet Said on Monday evening that OpenAI “should not be rushed” into its Defense Department deal.
Many saw OpenAI as bowing to the Pentagon’s will, leading Altman to claim that OpenAI would change the terms of the deal after the fact — a bizarre twist that likely won’t sit well with Trump’s military or the company’s already disillusioned customers.
Altman claimed the company would “amend our deal” to add a ban on “intentional tracking, surveillance, or monitoring of U.S. persons or citizens.”
“There are many things the technology is not yet ready for, and in many areas we still do not understand the tradeoffs required for security,” the CEO wrote. “We will work through these gradually, with the (War Department), with technical safeguards and other methods.”
It’s important to note that Altman’s tweet makes no mention of autonomous AI-enabled weapons systems, the other major issue that led to the rift between Anthropic and the Defense Department.
Whether or not this means that such weapons systems are on the table for OpenAI is unclear, but given Altman’s language, it’s certainly not out of the question.
It is also unclear whether the Defense Department would agree to these revised terms or whether it had originally agreed to accommodate OpenAI’s original terms and not Anthropic’s, as cnbc tells.
At the very least, Altman acknowledged that the optics of his eleventh-hour revision were disappointing.
He wrote, “We were really trying to tone things down and avoid a much worse outcome, but I think this just seemed opportunistic and lame.” “It’s a good learning experience for me because in the future we will be faced with high-risk decisions.”
Altman also called on the government not to designate Anthropic as a supply chain risk to national security. After Anthropic refused to sign the deal with the Pentagon, Defense Secretary Pete Hegseth said announced It stated on February 27 that “Effective immediately, any contractor, supplier, or partner that does business with the United States military may not conduct any business activity with Anthropic.”
“Their true purpose is unambiguous: to gain veto power over operational decisions of the United States military,” Hegseth said angrily. “this is unacceptable.”
Whether Altman’s latest mea culpa will meaningfully address OpenAI’s growing PR crisis is questionable. For many users, the damage has already been done. Moreover, it is not a one-horse race, ChatGPT is lagging behind every other leading AI company on the LLM leaderboard.
“People are excited and there are better products out there,” one Reddit user said. wrote. “It’s a recipe for disaster.”
But Anthropic’s Amodei may not be a knight in shining armor either. In form of wall street journal informed Over the weekend, the Defense Department selected targets in Iran using Anthropic’s cloud chatbot, highlighting the AI company’s pre-existing relationship with the military.
In other words, while Amodei pointed out CBS News carefully timed sunday interview Mass surveillance and autonomous weapons are “two red lines” that the company has had “from day one,” with the company still having signed a prior agreement with the Pentagon that allows it to use the cloud to carry out lethal attacks.
More information on the contract: Sam Altman is in damage control mode as ChatGPT users are mass unsubscribing because OpenAI is “training a war machine.”
