White House drafts guidance to bypass Anthropic's risk flag for new AI models: Report

White House reportedly drafts guidance to bypass Anthropic's safety flags, raising concerns over AI oversight and national security. | Contesto: cronaca

Punti chiave

• White House drafts guidance to bypass Anthropic's risk flag for new AI models: Report

Contesto

The White House has reportedly drafted guidance intended to circumvent safety protocols established by Anthropic, the artificial intelligence startup known for its rigorous risk assessment systems, according to a recent report. The move signals a growing tension between the Biden administration’s push to accelerate AI adoption for national security purposes and the tech industry’s efforts to self-regulate potentially dangerous applications. The guidance, details of which remain confidential, would allow federal agencies to override Anthropic’s built-in risk flags when deploying the company’s AI models. These flags are designed to prevent the use of AI in high-stakes scenarios such as autonomous weapons development or domestic surveillance—applications that Anthropic has publicly pledged to avoid. The development comes months after the startup faced a fallout with the Pentagon earlier this year, when it refused to remove guardrails against using its AI for autonomous weapons or domestic surveillance. Anthropic, founded by former OpenAI employees, has positioned itself as a safety-first company. Its models are trained to refuse certain requests that could lead to harm, and it has maintained a policy of not allowing its technology to be used for lethal autonomous systems. The company’s stance put it at odds with the Department of Defense, which has been exploring AI’s potential in military operations. The Pentagon’s interest in Anthropic’s models underscores the broader demand for advanced AI within the national security establishment, even as ethical concerns mount. The White House’s reported draft guidance would effectively create a workaround for federal agencies that find Anthropic’s safeguards too restrictive. Critics argue this could undermine the very safety measures the company has built its reputation on, and could set a precedent for other AI firms to weaken their own guardrails under government pressure. Supporters, however, contend that national security imperatives sometimes require flexibility, and that blanket restrictions could hamper legitimate defense and intelligence work. The broader implications of this guidance extend beyond Anthropic. If...

Lettura DEO

Decisione di validazione: publish

Risk score: 0.1

Il testo è stato ricostruito dai dati editoriali disponibili senza aggiungere fatti non presenti nel record sorgente.

Indicatore di affidabilità

Verificata — Alta confidenza. Fonti affidabili confermano la notizia.

Il sistema a semaforo

Ogni articolo su DEO include un indicatore di affidabilità:

• 🟢 Verificata — Alta confidenza. Fonti affidabili confermano la notizia.
• 🟡 In evoluzione — Confidenza moderata. Alcuni dettagli potrebbero ancora cambiare.
• 🔴 Contestata — Bassa confidenza. Fonti in conflitto o incertezze rilevanti.

Questo sistema esiste perché chi legge merita di sapere non solo cosa è successo, ma anche quanto la notizia è solida.

Categoria: cronaca