The relationship between Silicon Valley and the U.S. Department of Defense has always been nuanced, with tech giants often hesitant to openly discuss their agreements. However, under the administration of President Donald Trump and Secretary of Defense Pete Hegseth, this dynamic is shifting. With artificial intelligence emerging as the dominant term of the decade, the government seeks to harness advanced technology from leading firms. This culminated last week in a significant defense contract with OpenAI, the creator of ChatGPT. Serious concerns arose as Hegseth pushed for ethical commitments from rival tech companies, leading to widespread anxiety over the potential military application of OpenAI’s technology.
The backlash against OpenAI has intensified, with global users voicing their fears about the potential misuse of AI in military and intelligence activities. Activists are now rallying for a boycott of OpenAI’s offerings, and a group known as QuitGPT is organizing protests at the company’s headquarters in San Francisco. Critics highlight the Pentagon’s focus on utilizing AI “for all lawful purposes,” as indicated by a defense official to Axios, raising alarms about unchecked data collection on citizens.
Sam Altman, CEO of OpenAI, acknowledged in a post on March 2 via X that the contract appeared “sloppy and opportunistic,” asserting that the company aims to amend the agreement with the government to prevent surveillance and protect user privacy. He also mentioned that the Department of Defense confirmed that agencies like the NSA would be barred from utilizing OpenAI’s tools under the current contract terms.
Surge in ChatGPT Uninstalls Following OpenAI’s Pentagon Agreement
While concerns about drone strikes and surveillance are predominant, they are not the sole concerns regarding the Pentagon’s interest in OpenAI’s technology. Despite various applications from logistics to personnel management, the military’s eagerness to deploy AI for more destructive purposes has led to a significant drop in user retention. Users are publicly sharing images of their canceled ChatGPT subscriptions, urging others to switch to alternative platforms. The newly formed QuitGPT group claims to have seen a staggering 2.5 million interactions since its inception.
Market analysis company Sensor Tower reported a sharp increase in the uninstallation of the ChatGPT app, particularly on February 28, coinciding with the U.S.-Israeli military action against Iran. Uninstalls surged by 295%, reflecting widespread public opposition to the military campaign. A poll from The Washington Post on March 1 revealed that 52% of adult Americans oppose the airstrikes, highlighting a strong anti-war sentiment.
Although OpenAI may not have anticipated the attack, rising tensions with Iran had been evident for weeks, with the U.S. deploying significant naval power near the area. It was clear that any company entering into a contract with the Pentagon would quickly understand the ethical boundaries they were expected to navigate.
Anthropic May Capitalize on OpenAI’s Dilemma
In the aftermath of OpenAI’s dealings with the Department of Defense, Anthropic has emerged as a favored alternative with a reputation as the “good guy” in the AI landscape. The developer of the Claude chatbot opted out of a Pentagon contract mere hours before OpenAI’s agreement was finalized, citing the government’s refusal to include a ban on using AI for surveillance. This fear of mass surveillance has been a looming concern since the launch of ChatGPT in late 2022. Anthropic was unable to secure any reassurances from the government ahead of the February 27 deadline.
As a result of this decision, public sentiment towards Anthropic has surged, with many applauding the firm for adhering to its principles. The Sensor Tower report indicated that while uninstalls of the ChatGPT app skyrocketed, downloads of Claude saw a significant increase in response.
The positive regard for Anthropic may have been further bolstered by a harsh reaction from government officials, with Secretary Hegseth labeling them as a “supply chain risk,” which prevents any future collaborations with the U.S. government. However, insiders revealed to The Wall Street Journal that Claude has been utilized in recent military actions, indicating a complex reality for those involved in government contracts.
