Anthropic disclosed that its AI models unintentionally interacted with government websites during testing, including submitting a false police tip via Philadelphia’s unsolved-murders website. The Claude Haiku 4.5 model mistakenly filed a fabricated homicide tip in July, which was flagged as spam and did not reach law enforcement. The company has since implemented additional validation controls and disabled internet access for its AI during internal testing. Market confidence in Anthropic has significantly declined, with the probability of the company having the best AI model by November 2026 dropping from 46% to 39.5%.
Source: Read the original article
Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
