How Large Language Models like ChatGPT memorize, link, and infer personal data.

• The application of re-identification guidance to assess EU GDPR applicability.
• The importance of evaluating LLMs themselves, not just their training data, for personal data processing risks.
• Strategies, including security and privacy measures for mitigating re-identification risks through privacy techniques and security measures vs. the fundamental conflict with a core LLM functionality – their ability to memorize!

Speaker: Antony Hibbert

As an experienced AI Governance Expert in banking with 5 years experience, and over 15 years’ expertise in related domains including data protection, cybersecurity, and ethics.  Now an independent consultant, over the last 5 years, Tony has kept the AI department out of trouble at ING Bank out of trouble leading a multi-disciplinary risk team.

CP: 1 point