Western startups realign Chinese open-weight AI models to remove state censorship
Overview
Startups Hirundo and Lazarus AI say they edit the weights of Chinese open-weight models to remove state-aligned bias, censorship and dangerous knowledge before Western companies deploy them.
Hirundo co-founder Ben Luria says his company performs this realignment, while Lazarus AI says it removed mandatory censorship from Alibaba's Qwen models and deployed its ReAligned versions at about a dozen U.S. companies.
AI researcher Oriane Peter says a single Western-aligned safety standard cannot be met universally, challenging the premise of the effort.
Written by AI from the articles below · updated Nov 2, 6:59 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
- Rest of WorldNewsStartups realign Chinese open-weight AI models to Western safety and values standards
AIHirundo co-founder Ben Luria says his startup edits the weights of Chinese open-weight models to remove state-aligned bias, censorship, and dangerous knowledge before Western companies deploy them. Lazarus AI says it removed mandatory censorship from Alibaba's Qwen models and deployed its ReAligned models at about a dozen U.S. companies. AI researcher Oriane Peter says a universal Western-aligned safety standard is an impossible one to meet.
Heat Trend
- Current comparable heat
- 25
- Peak comparable heat
- 29Nov 2, 8:00 AM ET
- Change in past 24 hours
- –
The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.