Tags: censorship

The Trump administration unveiled a legislative framework aimed at creating a single, nationwide AI policy. The plan would centralize authority in Washington, preempting state AI regulations while emphasizing a light‑touch, innovation‑focused approach. It assigns greater responsibility for child safety to parents, calls on Congress to require platforms to add safeguards against sexual exploitation, and seeks to shield developers from state liability. Critics argue the proposal limits state experimentation and lacks clear enforcement mechanisms, while industry leaders praise the promise of a uniform national standard for startups. Weiterlesen

Researchers from Stanford and Princeton compared the responses of several Chinese and American large language models to politically sensitive questions. The study found that Chinese models refuse to answer a significantly larger share of these queries, provide shorter replies, and sometimes deliver inaccurate information. The authors suggest that manual fine‑tuning, rather than censored training data, drives much of this behavior. Additional work shows that extracting hidden instructions from Chinese models is difficult, highlighting the challenges of studying AI‑driven censorship in real time. Weiterlesen

X is contesting a Karnataka High Court ruling that would compel the platform to obey millions of takedown requests via the government‑run Sahyog portal, a system X describes as a censorship tool that bypasses judicial review. The company says the order threatens free expression and could expose platforms to criminal liability for non‑compliance. X’s appeal follows earlier disputes with Indian authorities, including challenges to block orders in 2024 and 2022 and a 2021 confrontation that saw officials threaten jail for Twitter employees. Weiterlesen

A massive data leak of over 500 GB from Geedge Networks exposed China’s export of deep‑packet‑inspection (DPI) technology—dubbed a “Great Firewall in a box”—to at least four authoritarian regimes. The leak includes source code, work logs and internal communications that detail how the DPI engine blocks VPN traffic, filters keywords, and shapes internet traffic. Researchers traced the hardware and software to Ethiopia, Myanmar, Kazakhstan and Pakistan, highlighting a shift from domestic control to a commercial censorship model. The export raises concerns for citizens, activists and journalists as VPNs struggle to bypass the new filters. Weiterlesen