AI Incident Database #357 · 2019-02-14
GPT-2 Reportedly Reproduced Personal Data from Its Training Data
OpenAI's GPT-2 reportedly memorized and reproduced portions of its training data, including personal information such as names, email addresses, social media handles, and phone numbers.
- Risk class (MIT)
- Compromise of privacy by obtaining, leaking or correctly inferring sensitive information
- Reports
- 3
- Developers and deployers named
- 1
On ShieldMap: OpenAI (named by AIID as a developer or deployer).
Reports
- What happens when your massive text-generating neural net starts spitting out people's phone numbers? If you're OpenAI, you create a filter · theregister.com, 2021-03-18
- Does GPT-2 Know Your Phone Number? · bair.berkeley.edu, 2020-12-20
- Extracting Training Data from Large Language Models · arxiv.org, 2020-12-02
Title and description from the AI Incident Database (incidentdatabase.ai), snapshot of 2026-11-16, shared under CC BY-SA 4.0. Selected incidents; descriptions shortened; harmed parties and report text left out. The description is AIID's summary of what was reported, not ShieldPi's finding.