Anthropic has spoken more about the risks of AI this year than its rival OpenAI, with critics suggesting such warnings might have triggered a US ban on foreign access to their newest models.

In an analysis by FT, it was found that five in every 1,000 words used by Anthropic related to risk or restrictions, compared to just one-sixth of this figure for OpenAI. This disparity has sparked debate and concern among technologists and policymakers about the balance between caution and regulation.

Yann LeCun, a pioneer in AI, criticised Anthropic’s chief Dario Amodei for his “ridiculous fear-mongering,” which he believes has led to this ban. This event is seen as an early test of how the US intends to oversee powerful new AI models.

The incident highlights the tension between companies warning about AI risks and facing potential backlash, with some questioning whether such warnings could inadvertently lead to overly restrictive measures.