Artificial Intelligence in the Hands of the Pentagon: Anthropic and OpenAI in the Shadow of Attacks in Iran

The situation in the artificial intelligence market has intensified sharply following revelations by the Washington Post regarding the use of generative technologies for military purposes. According to available information, a model developed by Anthropic was used by the US Department of Defense to identify strategic targets during airstrikes aimed at facilities in Iran. This event sparked a wave of controversy, as Anthropic had previously positioned itself as a company prioritizing ethics and safety over pure profit. Despite earlier tensions and a brief severing of ties with the American administration, the company has returned to the negotiating table, seeking to establish rules for cooperation with the federal government. At the same time, OpenAI, the creator of ChatGPT, finds itself at a similar turning point. Sam Altman admitted that the company does not have full control over how the Pentagon uses the tools provided to it. Signals are emerging about ambitious plans to extend this cooperation to the classified systems of the North Atlantic Alliance. These plans have met with fierce public resistance – in San Francisco, massive protests have erupted against military contracts for tech companies. Critics fear that automating decision-making processes on the battlefield could lead to an uncontrolled escalation of armed conflicts. Since the end of the Cold War, the Pentagon has sought to maintain technological superiority, but only the development of large language models in the last decade has enabled the real-time analysis of massive intelligence data sets. Against the backdrop of these events, a significant shift has occurred in the financial sector. Jensen Huang, CEO of Nvidia, announced the suspension of a planned investment in OpenAI, which could have reached an astronomical sum of 100 billion dollars. The official reason is a change in investment strategy, but analysts suggest Nvidia fears reputational risk associated with the military use of AI and growing competition from Broadcom, which forecasts a sharp increase in sales of chips dedicated to artificial intelligence. Additionally, Nvidia has halted deliveries of advanced H200 processors to China, shifting production capacity to newer units like Vera Rubin, which fits into the broader context of US technology sanctions. „Sam Altman admits OpenAI can't control Pentagon's use of AI” — Sam Altman An unexpected phenomenon is the reaction of the consumer market. Despite ethical controversies, the Claude application recorded a record surge in popularity, which some commentators ironically attribute to 'anti-advertising' from Donald Trump and loud debates about its power. This situation illustrates that the line between civilian software and strategic weaponry is ultimately blurring, posing a challenge for legislators to create new regulatory frameworks for technology capable of planning military strikes.

Perspektywy mediów: Liberal media emphasize the ethical dangers and social protests against using AI for killing. Conservative media stress the necessity of building US technological superiority over Iran and China.