Amazon Security Research Reportedly Led to White House Ban on Anthropic AI Access
An export control directive has resulted in AI firm Anthropic restricting access to its Fable 5 and Mythos 5 models. This decision was reportedly influenced by cybersecurity research conducted by Amazon and discussions between Amazon CEO Andy Jassy and the White House. According to a Wall Street Journal report, Amazon's research indicated that Fable 5 could generate information useful for cyberattacks when given specific prompts.
An export control directive has led to AI firm Anthropic restricting access to its Fable 5 and Mythos 5 models. This development was reportedly influenced by cybersecurity research from Amazon and discussions held between Amazon CEO Andy Jassy and the White House.
A report from the Wall Street Journal indicates that Amazon's research paper claimed Fable 5 was capable of producing information that could be leveraged in cyberattacks. This capability was reportedly demonstrated through a series of specific prompts.
Following CEO Jassy's sharing of Amazon's findings with the government, a decision was made to block the use of these AI models by foreign nationals. Amazon has not yet responded to requests for comment regarding these reports.
According to The Verge, which cited the Wall Street Journal report, these events unfolded as a result of the security concerns raised.
Advertisement
AdSense slot • inline
