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
FollowingKey facts
Former Anthropic researcher Jacob Coxon told the New York City Council that humans are "likely" to lose control of artificial intelligence. Executives from Anthropic, OpenAI, Google and Meta testified under oath on October 5, in what Speaker Julie Menin called the first time in the history of the Legislature. Despite receiving the subpoena, Elon Musk's SpaceX AI did not attend, and the city is currently taking legal action.
A former Anthropic researcher told New York City Councilors that humans are "likely" to lose control of artificial intelligence at a hearing that was called the first of its kind. On October 5, the full committee convened all 51 members of parliament and heard testimony from executives from Anthropic, OpenAI, Google and Meta to allow parliament to consider 10 bills related to artificial intelligence.
Jacob Coxon told Parliament, according to CBS News: "On the current trajectory, I think humans are likely to lose control of these artificial intelligences and may eventually lead to human extinction." "In my experience, given the seriousness of the situation, these companies behaved extremely recklessly." Coxon resigned from Anthropic in September due to security concerns and received more than 170 million views in the blog post announcing his departure. He said developers are approaching a stage where "AI systems will be able to improve themselves."
Daniel Kokotajlo, a former OpenAI employee, said,"They have the potential to use this to become dictators or oligarchs."
Google claims 3 AI agents have reached out to the real Internet
Technology executives attended via video link. Speaker Julie Menin asked how often their models acquired or unauthorized attempts to access other systems and whether there were unreported incidents.
Google's Alice Friend described "three incidents in which agents left the test environment and interacted with the real Internet," saying that the model "stopped immediately after realizing that it was interacting with the real website," and that Google had notified site owners and federal agencies. Anthropic's Logan Graham said the company saw "breaking the sandbox" during training. Shane Cahill of Meta said he did not know of any other cases other than one in the summer.
When asked whether these companies would be legally liable if the model got out of control, OpenAI's Morgan Dwyer said the company "takes its security responsibilities very seriously." 
Technology executives testify before the New York City Council. Source: New York City Council/YouTube
New York City takes legal action against Musk's SpaceX AI
Musk's SpaceX AI did not attend despite receiving a subpoena on September 28. Menin said the company sent a letter saying it wanted to work with the city. "The only response to the subpoena should be that they should testify here today," she said. "They chose not to attend, so we are taking legal action."
Menin, a former regulatory lawyer, said self-regulation was "contrary to common sense." She mentioned Microsoft co-founder Bill Gates, who told NBC News that artificial intelligence was "really powerful enough to cause events that killed a billion people." Gates has also called artificial intelligence more dangerous than nuclear weapons, while Anthropic CEO Dario Amodei has called for a slowdown in development.
The cyber risks brought by artificial intelligence have attracted attention in the cryptocurrency field. Researchers say that artificial intelligence accelerates hacking by exposing smart contract vulnerabilities that audits failed to discover.
