Exchange Ranking
Top Exchanges
24h Volume Ranking
Popularity Ranking
Exchange BTC Balance
Proof of Reserves
Decentralized Exchanges
Funding Rate
Funding Heatmap
Liquidation Data
Max Pain
Long/Short Ratio
Whale L/S Ratio
Binance/Okex/Huobi L/S
Bitfinex Margin L/S
ETF Tracker
Solana ETF
XRP ETF
Hong Kong ETF
Bitcoin Treasuries
Crypto Reversal
Ethereum Reserves
HyperLiquid Wallet Analysis
Hyperliquid Whale Watch
Large Transactions
On-chain Movement
Bitcoin ROI
Stablecoin Market Cap
Options Analysis
News
Articles
Economic Calendar
Features
Wallet
Contract Calculator
Security
Collections
Watchlist
FollowingThree conversations with Anthropic's Claude triggered police intervention between August and September
Between August 11 and September 30, a total of three conversations involving Anthropic's artificial intelligence model Claude were submitted to the police. Two of those users are currently facing felony charges.
The third person involved was a man from San Francisco who allegedly threatened CEO Dario Amodi, but the man was not detained.
The San Antonio arrest and AR-15 shooting threats against Amodi start August controversy
The first incident occurred on August 11 in San Antonio, Texas. According to an arrest affidavit, 22-year-old Nathaniel Michael Carrasco said in a conversation with Anthropic's artificial intelligence that he wanted to buy a gun and shoot students at Serna Elementary School.
The FBI's National Threats Action Division notified local police on August 28. Law enforcement officials linked Carrasco's phone number, IP traffic and address to the account by sending emergency disclosure requests to T-Mobile, Charter Communications and Google.
The school's principal said Carrasco was arrested on August 30 on a third-degree felony charge of making a terrorist threat that caused public fear of serious bodily harm. She pointed out that schools learned about specific threats against their campuses through news reports.
On August 14, three days after the San Antonio conversation, San Francisco police reported that a man claimed to have disclosed to Claude that he had purchased an AR-15 rifle and had "targeted Amodi." The following week, on Tuesday, Anthropic notified the San Francisco Police Department (SFPD) of the incident.
Police confirmed that the man was not in the company's office at 500 Howard Street. An Anthropic spokesman said the company had blocked the account and referred the case to law enforcement, saying it was a reflection of the normal operation of security mechanisms. As of early September, the man had not been arrested or charged.
The Anthropic review team sent a message to the Lee County Sheriff's Office on September 26
Anthropic's privacy policy effective September 10 stipulates that data can be shared with law enforcement if the company believes it is necessary to prevent serious harm in good faith. Its Consumer Service Terms reserve the right to report user inputs, outputs or actions to law enforcement at the company's sole discretion.
Arrest reports show that Carli Michelle Heller of Bonita Springs, Florida, wrote on September 26 that she would go to the Lee County Sheriff's Office to "kill people." Investigators said she added the next day that she had obtained new guns.
The report states that these remarks were flagged by Anthropic's surveillance system. Due to its severity, the manual review team referred it to law enforcement for processing.
警员在海勒家中将其拘留，过程未发生意外。她因涉嫌作出书面暴力威胁而面临指控，并于11月出庭受审。警长卡明·马塞诺（Carmine Marceno）表示，她后来形容使用AI如同写日记一般。
马塞诺在声明中表示：“人工智能是一种强大的工具，就像任何技术一样，它也可能被滥用。”
背景信息：Anthropic的安全措施演进
此前有报道指出，Anthropic于今年4月开始要求部分功能用户提供政府签发的身份证件及实时自拍验证。同年8月，该公司宣布调整Claude的记忆设置，除非用户主动选择加入，否则系统将跳过与健康、政治及性别认同相关的话题。
