MITRE Launches AI Incident Sharing Initiative

MITRE’s Middle for Menace-Knowledgeable Protection introduced the launch of the AI Incident Sharing initiative this week, a collaboration with greater than 15 firms to extend group information of threats and defenses for AI-enabled methods

The incident sharing initiative falls underneath the purview of the middle’s Safe AI challenge, and goals to allow fast and safe collaboration on threats, assaults and accidents involving AI-enabled methods. It expands the attain of the MITRE ATLAS group information base, which has been accumulating and characterizing information on anonymized incidents for 2 years. Below this initiative, a group of collaborators will obtain protected and anonymized information on real-world AI incidents. 

Incidents could be submitted through web (at https://ai-incidents.mitre.org/) by anybody. Submitting organizations will likely be thought-about for membership with the purpose of enabling data-driven danger intelligence and evaluation at scale. 

Safe AI additionally prolonged the ATLAS menace framework to include data on the generative AI-enabled system menace panorama, including a number of new generative AI-focused case research and assault strategies, in addition to new strategies to mitigate assaults on these methods. In November 2023, in collaboration with Microsoft, MITRE launched updates to the ATLAS information base centered on generative AI. 

“Standardized and speedy data sharing about incidents will permit your complete group to enhance the collective protection of such methods and mitigate exterior harms,” stated Douglas Robbins, vice chairman, MITRE Labs, in a statement.

MITRE operates an analogous information-sharing public non-public partnership with the Aviation Security Data Evaluation and Sharing database for sharing information and security data to determine and forestall hazards in aviation.

Collaborators on Safe AI span industries, with representatives from monetary providers, expertise, and healthcare. The record contains AttackIQ, BlueRock, booz Allen Hamilton, CATO Networks, Citigroup, Cloud Safety Alliance, CrowdStrike, FS-ISAC, Fujitsu, HCA Healthcare, HiddenLayer, Intel, JPMorgan Chase Financial institution, Microsoft, Customary Chartered, and Verizon Enterprise.