Google Offers Up to $20,000 in New AI Bug Bounty Program

Google has unveiled a refreshed AI Vulnerability Reward Program (VRP), expanding its scope and merging rewards for abuse and security issues into one table, with the highest payouts reaching $20,000 for attacks that modify a victim’s account or data in flagship products. According to Google explains, rewards for similar attacks in standard products go up to $15,000, while the most a researcher can earn for exfiltration of data from flagship or standard products is $15,000 and $10,000 for the “other” tier.

The programme notes that prompt injections, jailbreaks, and alignment issues are not within scope, though researchers are urged to report content-related issues via in-product reporting channels. Attacks described as enabling persistent DoS or data leakage, or exfiltration of model parameters, are among those included in scope, and Google has three product tiers: flagship, standard, and other.

A unified rewards panel will review all payouts and issue the highest reward across the abuse and security tables, with the rules detailed on the programme’s page.