Major Security Vulnerabilities in AI Systems ChatGPT and OpenClaw Lead to Data Breach Risks and Real-World Harm

Thumbnail Image

The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.

OpenAI patched a ChatGPT vulnerability that could leak user data via prompt injection and DNS queries, averting potential privacy breaches. Separately, 360's AI agents discovered critical flaws in OpenClaw, exposing over 170,000 instances globally to data theft and attacks, with actual cases of financial loss reported. Both incidents highlight significant AI security risks.[AI generated]

Why's our monitor labelling this an incident or hazard?

Langflow is an AI system development tool for large language models, thus clearly involving AI. The vulnerability allows attackers to remotely execute arbitrary code, which has been confirmed to be exploited in the wild, causing direct harm through unauthorized access and potential data breaches. This meets the criteria for an AI Incident because the AI system's malfunction (security vulnerability) has directly led to harm. The event is not merely a potential risk or a complementary update but a confirmed exploitation causing harm.[AI generated]
AI principles
Privacy & data governanceRobustness & digital security

Industries
Digital security

Affected stakeholders
ConsumersBusiness

Harm types
Human or fundamental rightsEconomic/Property

AI system task:
Interaction support/chatbots


Articles about this incident or hazard