French watchdog to probe DeepSeek’s AI data practices

France’s privacy regulator CNIL announced plans to investigate Chinese AI startup DeepSeek’s operations and data protection practices, following similar probes launched by Irish and Italian authorities.

Key development: The French data protection authority (CNIL) will examine DeepSeek’s AI system to understand its functionality and assess potential privacy risks for users.

• DeepSeek recently gained attention by claiming it trained its DeepSeek-V3 model for less than $6 million using Nvidia H800 chips
• CNIL will specifically question the company about its chatbot operations and data protection measures
• The investigation joins similar inquiries launched by privacy regulators in Italy and Ireland

Regulatory context: The European Union maintains some of the world’s strictest data protection regulations through GDPR and newly established AI rules.

• Companies violating GDPR can face fines of up to 4% of their global turnover
• The EU’s new AI rules impose strict transparency requirements on high-risk AI systems
• Violations of AI regulations can result in fines ranging from 7.5 million euros (1.5% of turnover) to 35 million euros (7% of turnover)

European regulatory landscape: CNIL has established itself as one of Europe’s most active privacy watchdogs.

• The French regulator has previously imposed fines on major tech companies including Google and Meta Platforms
• Ireland’s Data Protection Commission is seeking information about DeepSeek’s data processing related to Irish users
• Italy’s privacy authority has also launched an investigation into DeepSeek’s personal data practices

Looking ahead: As Chinese AI companies expand their global presence, they face increasing scrutiny from European regulators concerned with data protection and AI safety standards.