Slack AI Vulnerability Exposes Private Channel Data

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.

Salesforce's Slack AI had a vulnerability allowing attackers to access private channel data via prompt injection, a flaw in its large language model. Security firm PromptArmor discovered this issue, which could lead to unauthorized data access and phishing attacks. Slack has since patched the flaw to protect user privacy.[AI generated]

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

Slack AI’s flaw enabled a novel method for hackers to exploit the chatbot and deliver malware, creating a clear risk of data breaches and harm. Because the vulnerability was discovered and patched before documented damage occurred, this is a plausible future harm rather than a realized incident.[AI generated]
AI principles
Privacy & data governanceRobustness & digital securityRespect of human rightsSafetyAccountability
Industries
IT infrastructure and hostingDigital securityBusiness processes and support services
Affected stakeholders
WorkersBusiness
Harm types
Human or fundamental rightsEconomic/PropertyReputational
Severity
AI hazard
Business function:
Citizen/customer service
AI system task:
Interaction support/chatbotsContent generation

Articles about this incident or hazard