Despite the issues, Amazon downplayed the significance of the internal discussions, stating that no security issue was identified as a result of the feedback. The company also emphasized that Q, which is currently available in a free preview, is more secure than consumer-grade tools like ChatGPT. However, an internal document acknowledged that Q can return harmful or inappropriate responses, including out-of-date security information that could put customer accounts at risk.
Key takeaways:
- Amazon's AI chatbot Q is reportedly experiencing severe issues, including "hallucinations" and leaking confidential data, according to leaked documents obtained by Platformer.
- The incident has been marked as "sev 2," which is serious enough to require engineers to work overnight and through the weekend to resolve it.
- Despite these issues, Amazon is downplaying the significance of the problems, stating that no security issue was identified as a result of the feedback received.
- Q was initially presented as a secure enterprise-software version of ChatGPT, capable of answering developers’ questions about AWS, editing source code, and citing sources.