Researchers discovered 18 malicious AI browser extensions masquerading as productivity tools that deliver remote access trojans, meddler-in-the-middle attacks, and infostealers. These extensions exploit the rise of generative AI to target prompts, user behavior, and browser sessions through API interception, passive DOM observation, traffic proxying, and HTTPS response decryption. Examples include extensions that surveil emails during composition, intercept ChatGPT prompts, and exfiltrate passwords. Multiple samples contained AI-generated code indicating threat actors employed large language models to accelerate production. Google removed or issued warnings for all 18 reported extensions. These malicious tools specifically target sensitive data including AI API keys, authentication credentials, email content, and proprietary session information by exploiting user trust in AI-branded applications.