Amazon’s New AI Chatbot, Q, Suffers “Severe Hallucinations” and Leaks Confidential Data
Earlier this week, Amazon made an exciting announcement, revealing its entry into the AI chatbot competition with the launch of its own chatbot named Q.
Amazon Web Services (AWS) clarified that Q is not intended to directly rival OpenAI’s powerful AI tools like ChatGPT. Instead, it will go head to head against Microsoft’s AI developer assistant, Copilot. AWS CEO Adam Selipsky stated that Amazon Q will serve as the AI assistant for businesses, allowing them to query their own business data. For instance, employees can ask Q about relevant company information and sift through vast amounts of data.
However, leaked documents obtained by Platformer have shed light on a concerning issue. Q has reportedly started experiencing “severe hallucinations” and, unfortunately, has been “leaking confidential data.” This leak was not a minor incident, as AWS engineers had to work tirelessly over the weekend to address the problem labeled as “sev 2.” Shockingly, this error exposed the locations of AWS data centers and unreleased features.
Data leaks are never a favorable situation, particularly for companies striving to maintain the utmost security for their valuable information. Such leaks can have detrimental effects on profits, lead to intellectual property loss, unveil insider secrets benefiting competitors, and more. With Q already leaking information about AWS, this incident deals a significant blow to the confidence that companies may have in this technology.
It remains to be seen how Amazon will address these issues and regain trust in its AI chatbot.