bit.fanBit.FanWeb3 News · Betting Reviews · Market Insights
‹ BackNewsAI compliance

AI compliance

Anthropic
2026-08-20 03:25:26

OKX limits Claude access for Hong Kong staff as Anthropic geo rules trigger compliance issues

Anthropic’s location-based access rules for Claude have created a new compliance problem for multinational companies, and OKX is now among the firms affected. According to Bloomberg, the crypto exchange has begun restricting Hong Kong-based employees from using Claude because of Anthropic’s geographic access policy. OKX founder and CEO Star Xu later said the company is working with Anthropic to resolve the issue and restore account access, while also tightening internal large language model permissions so only employees in approved regions can use Claude. Xu’s response also offered a rare look at OKX’s internal AI spending and workflow. He said the company spends between $6 million and $8 million a month across several leading foundation model providers. OKX has built an internal AI development platform called Oli rather than relying on separate subscriptions to services such as ChatGPT or Claude. The platform combines multiple frontier models in one environment for product managers and engineers. Xu added that more than 75% of OKX’s weekly product releases and pull requests are now completed with help from Oli, across more than 1,000 releases and pull requests each week.

1320
OKX limits Claude access for Hong Kong staff as Anthropic geo rules trigger compliance issues
People’s Publ
2026-08-02 12:34:03

Chinese police university researchers unveil AI framework to detect illegal crypto transactions

Researchers at the People’s Public Security University of China have developed an AI framework that can detect illegal cryptocurrency transactions with overall accuracy close to 90%, according to a report cited by ChainCatcher. The study was published in the Chinese peer-reviewed journal Intelligence Magazine. Sun Jingchao, the corresponding author and a researcher focused on criminal investigation and cybersecurity, said the work offers an accurate, scalable and interpretable solution for identifying illicit crypto activity, while also giving regulators a new technical route for tackling economic crimes tied to digital assets. The framework uses a memory module and a large language model to identify patterns linked to crypto’s anonymity and cross-border nature, with the aim of tracing illegal activity such as money laundering. Its release comes as Chinese authorities continue to intensify enforcement against financial crimes involving cryptocurrencies. In March this year, China’s Supreme People’s Procuratorate said prosecutors had charged 3,259 people in 2025 in money-laundering cases involving virtual currencies and underground banks.

1800
Chinese police university researchers unveil AI framework to detect illegal crypto transactions