Microsoft has introduced its AI bug-bounty program to encourage researchers worldwide to find vulnerabilities throughout the Bing generative AI chatbot and AI integrations. Bounty rewards will vary from $2,000 to $15,000 for certified submissions.
Eligible members should be not less than 14 years outdated, with permission from a authorized guardian if they’re a minor, and a person researcher. Ought to a participant be a public sector worker, the bounty award should go to the general public sector group and be signed by an lawyer or govt accountable for its ethics insurance policies.
The scope of the bounty program extends to AI-powered Bing on bing.com, AI-powered Bing integration in Microsoft Edge, AI-powered Bing integration within the Microsoft Begin app, and AI-powered Bing integration within the Skype Cellular app. Any vulnerabilities present in these integrations are certified for submission and are eligible to win a reward.
Microsoft said that the objective of this system is to uncover vulnerabilities which have a major influence on the safety of its prospects throughout the AI-powered “Bing expertise.” When submitting a vulnerability, researchers should make sure that it has not been beforehand reported, is of important or essential severity as per the Microsoft Vulnerability Severity Classification for AI Methods, and is reproducible on the newest model of the product with clear steps as to methods to reproduce the vulnerability.
Additional instructions as to methods to get began and enter a submission; particular data on various kinds of vulnerabilities and the winnings they’ve the potential to earn; analysis guidelines of engagement; and phrases and situations, are listed on Microsoft’s web site.
