Tensions between two of the world’s leading artificial intelligence (AI) companies seem to be intensifying. Dario Amodei, chief executive of Anthropic, has criticised rival OpenAI and its chief executive, Sam Altman over the company’s recent agreement with the United States Department of War.
In a message to employees, Amodei said the company’s relationship with the American government had frayed because it was unwilling to cosy up to the administration and give “dictator-style” praise to Donald Trump, unlike other AI companies (read as: OpenAI)
Amodei ridiculed OpenAI’s approach to working with the US military, describing the company’s position as “safety theatre”, according to a report by The Information.
“The main reason [OpenAI] accepted [the DoD’s deal] and we did not is that they cared about placating employees, and we actually cared about preventing abuses,” Amodei wrote.
Also Read: ETtech Explainer: Anthropic’s rapid rise, Pentagon standoff and everything in between
The Anthropic CEO’s remarks followed the collapse of negotiations between the company and the Pentagon over the military’s request for unrestricted access to its technology. Anthropic, which already had a $200 million contract with the military, insisted that the DoD affirm that it would not use the company’s AI to enable domestic mass surveillance or autonomous weaponry.
Instead, the Pentagon reached a separate deal with OpenAI. Altman said the agreement would include protections relating to the same ethical boundaries that Anthropic had insisted upon.
Amodei disputed this claim in his staff letter, accusing OpenAI of misrepresenting the situation. In the message, he described the company’s public statements as “straight up lies” and argued that Altman was wrongly “presenting himself as a peacemaker and dealmaker.”
Anthropic’s concern largely centred on the Pentagon’s demand that the company’s AI systems be available for “any lawful use”. Instead, the company wanted stronger safeguards written directly into the agreement.
OpenAI, meanwhile, said its contract allows the use of its AI systems for “all lawful purposes”. However, in a blog post, the company explained that discussions with defence officials clarified that domestic mass surveillance would not fall under that category.
“It was clear in our interaction that the DoW considers mass domestic surveillance illegal and was not planning to use it for this purpose.” OpenAI’s blog post stated. “We ensured that the fact that it is not covered under lawful use was made explicit in our contract.”
Some critics argue that laws can change over time, meaning activities currently considered illegal could later become permissible.
Adding to the debate, a report by Bloomberg said Altman told employees that OpenAI does not ultimately control how the defence department uses its technology, which seems to contradict the company’s blog post. According to the report, he even suggested that Anthropic’s insistence on the matter may have contributed to the earlier tensions with the Pentagon.
Amodei also addressed public reaction to the dispute in his internal message, claiming that attempts to shape the narrative were not convincing to many observers.
“I think this attempted spin/gaslighting is not working very well on the general public or the media, where people mostly see OpenAI’s deal with the DoW as sketchy or suspicious, and see us as the heroes (we’re #2 in the App Store now!),” Amodei wrote to his staff. “It is working on some Twitter morons, which doesn’t matter, but my main worry is how to make sure it doesn’t work on OpenAI employees.”
Following OpenAI’s defence deal, uninstallations of its chatbot surged 295% day over day on Saturday, with downloads of ChatGPT falling sharply as well. In contrast, Anthropic’s Claude climbed to the number one spot of free app rankings on the US Apple App Store during the weekend
In a message to employees, Amodei said the company’s relationship with the American government had frayed because it was unwilling to cosy up to the administration and give “dictator-style” praise to Donald Trump, unlike other AI companies (read as: OpenAI)
Amodei ridiculed OpenAI’s approach to working with the US military, describing the company’s position as “safety theatre”, according to a report by The Information.
“The main reason [OpenAI] accepted [the DoD’s deal] and we did not is that they cared about placating employees, and we actually cared about preventing abuses,” Amodei wrote.
Also Read: ETtech Explainer: Anthropic’s rapid rise, Pentagon standoff and everything in between
The Anthropic CEO’s remarks followed the collapse of negotiations between the company and the Pentagon over the military’s request for unrestricted access to its technology. Anthropic, which already had a $200 million contract with the military, insisted that the DoD affirm that it would not use the company’s AI to enable domestic mass surveillance or autonomous weaponry.
Instead, the Pentagon reached a separate deal with OpenAI. Altman said the agreement would include protections relating to the same ethical boundaries that Anthropic had insisted upon.
Amodei disputed this claim in his staff letter, accusing OpenAI of misrepresenting the situation. In the message, he described the company’s public statements as “straight up lies” and argued that Altman was wrongly “presenting himself as a peacemaker and dealmaker.”
Anthropic’s concern largely centred on the Pentagon’s demand that the company’s AI systems be available for “any lawful use”. Instead, the company wanted stronger safeguards written directly into the agreement.
OpenAI, meanwhile, said its contract allows the use of its AI systems for “all lawful purposes”. However, in a blog post, the company explained that discussions with defence officials clarified that domestic mass surveillance would not fall under that category.
“It was clear in our interaction that the DoW considers mass domestic surveillance illegal and was not planning to use it for this purpose.” OpenAI’s blog post stated. “We ensured that the fact that it is not covered under lawful use was made explicit in our contract.”
Some critics argue that laws can change over time, meaning activities currently considered illegal could later become permissible.
Adding to the debate, a report by Bloomberg said Altman told employees that OpenAI does not ultimately control how the defence department uses its technology, which seems to contradict the company’s blog post. According to the report, he even suggested that Anthropic’s insistence on the matter may have contributed to the earlier tensions with the Pentagon.
Amodei also addressed public reaction to the dispute in his internal message, claiming that attempts to shape the narrative were not convincing to many observers.
“I think this attempted spin/gaslighting is not working very well on the general public or the media, where people mostly see OpenAI’s deal with the DoW as sketchy or suspicious, and see us as the heroes (we’re #2 in the App Store now!),” Amodei wrote to his staff. “It is working on some Twitter morons, which doesn’t matter, but my main worry is how to make sure it doesn’t work on OpenAI employees.”
Following OpenAI’s defence deal, uninstallations of its chatbot surged 295% day over day on Saturday, with downloads of ChatGPT falling sharply as well. In contrast, Anthropic’s Claude climbed to the number one spot of free app rankings on the US Apple App Store during the weekend
