Just a short while after Amazon introduced its new AI chatbot, Q, there are concerns raised by some of the company’s employees. They’re warning about potential serious issues like “severe hallucinations” and the possibility of sensitive information being exposed during the public preview of this new tool.
Apparently, Q has the capability to reveal quite sensitive data, including the locations of AWS data centers, internal discount programs, and features that haven’t been released yet. This info comes from Amazon employees as reported by Platformer.
The problems seem pretty serious, enough so that engineers are getting called in at odd hours and having to put in extra hours over the weekends. According to internal documents, some of these hallucinations are alarming enough to cause potential health issues within the legal department, even heart attacks.
However, Amazon is pushing back against these concerns. They’ve stated through a spokesperson that they regularly gather feedback from employees using internal channels and ticket systems, and based on that feedback, they haven’t found any security problems.
“We’re grateful for the feedback we’ve received so far and we’re committed to improving Q as it moves from a preview to a fully available product,” the spokesperson mentioned. Later on, they added that according to their assessment, Q hasn’t leaked any confidential information.
Q is primarily aimed at businesses to assist their employees in various tasks such as summarizing documents, handling tickets, answering questions, and managing actions related to source codes and company data.
Adam Selipsky, CEO of Amazon Web Services, had recently compared Q to other AI chatbots, saying it’s safer than those targeted at general consumers like ChatGPT, which faced bans from enterprises due to worries about security and privacy. Overcoming these concerns is crucial for Q to find its place and succeed in the business environment.
