A newly published academic paper has revealed a critical vulnerability in AI-powered deep-research systems, including those underpinning commercial tools like OpenAI’s Deep Research and Google’s Gemini Deep Research, that allows a single short Reddit comment to manipulate the reports these agents...