Last seen October 9, 2025

AI Security Incident

Evidence indicates that the affected technology is affected by a security issue.

Technical Severity
Low severity
Lifecycle Status

STABLE

What Happened

Evidence indicates that the affected technology is affected by a security issue.

Why This Matters

Current evidence identifies a security issue involving the affected technology, but does not yet support a more specific impact claim.

Recommended Action

No confirmed vendor remediation is available in the current evidence. Confirm whether the affected technology is present in your environment and review the affected configuration.

Exposure

Recommended Response
First Seen

Oct 09, 2025 05:30

Last Seen

Oct 09, 2025 05:30

Exploitation status: UNKNOWN

Primary entities:

Data Poisoning

Timeline

  • Incident first seen
    Oct 09, 2025 05:30

    BugSkan first recorded this incident.

  • A small number of samples can poison LLMs of any size - Anthropic
    Oct 09, 2025 05:30

    anthropic.com · Vulnerability

Sources

A small number of samples can poison LLMs of any size - Anthropic

anthropic.com · Oct 09, 2025 05:30

Researchers demonstrated that as few as 250 poisoned documents can create a backdoor vulnerability in large language models, irrespective of model size or training data volume. This data poisoning technique, which can induce denial-of-service or potentially facilitate data exfiltration, challenges prior assumptions about the required scale of malicious training data during pretraining.

Open publisher source

Watchlist Match

Want personalized relevance?

Create an account to see which incidents overlap with the technologies you monitor.

← Back to incident intelligence