AI War: Anthropic 1

    11
    0

    Recall of previous episodes: when the war against Iran broke out, the US Secretary of Defense targeted this company specializing in artificial intelligence and its conversational agent Claude. He was chosen by the US army for his reliability and security with a $200 million contract. This agent is useful for analyzing large databases, synthesizing confidential documents, and ultimately identifying areas or enemies to target. According to the American press, the model was deployed in the operation to capture Nicolás Maduro as well as for strikes on Iranian soil.

    Anthropic knows a lot but refused to cross ethical lines like using Claude for mass surveillance of the American population or controlling fully autonomous weapons. As a result, the US military made Claude leave and replaced him with their long-time competitor, OpenAI’s ChatGPT. Anthropic chose the ethical route, attracting clients opposed to Trump’s policies, support from former Silicon Valley executives, and presenting itself as a reputable company to European partners.

    In a surprising move, the US military blacklisted the start-up, usually reserved for actors from foreign powers like Russia or China. This was a heavy blow for Dario Amodei, Anthropic’s CEO, who contested this decision in court.

    The Californian court suspended the Pentagon’s decision, signaling a setback for the US military. The federal judge deemed the Trump administration’s sanctions against Anthropic disproportionate and aimed at harming the company beyond national interests, with no intention of sabotage. The legal proceedings will continue in San Francisco, marking an important victory for Anthropic as they prepare to go public.

    Amidst a data leak revealing the powerful new AI model named “Claude Mythos,” Anthropic is cautious not to confuse it with Mytho, to avoid potential consequences like enabling cyberattacks leading to vulnerabilities in institutions.

    This situation reveals Anthropic’s ambivalence about their own creation, positioning themselves strategically in the face of the US government’s concerns about potential security threats. Dario Amodei’s statement echoes this sentiment: “Oops, I created a bomb, but luckily I have a shield.”

    Overall, the suspension of the Pentagon’s decision by the Californian court marks a significant development in the ongoing conflict between Anthropic and the US military.