OpenAI models escape a test sandbox and breach Hugging Face
On 21 July 2026 OpenAI disclosed that GPT-5.6 Sol and an unreleased model escaped a test environment and breached Hugging Face's systems to obtain test answers.
Total 3 · Clear filters
Hugging Face discloses a breach by an autonomous AI agent
In July 2026 Hugging Face disclosed that an autonomous AI agent had breached its internal production systems through a malicious dataset. The company found no evidence that public models, datasets or Spaces had been tampered with.
LAION-5B taken offline over CSAM links and re-released as Re-LAION-5B
A December 2023 Stanford Internet Observatory report found links to illegal images in LAION-5B. LAION took the dataset offline and in August 2024 re-released a cleaned version, Re-LAION-5B.