Research indicates that depending on AI may transform uncertainty into misplaced assurance.

      AI is making it increasingly difficult for individuals to admit "I don't know."

      You might assume that receiving incorrect advice would lead to more caution. However, new research indicates the contrary when that advice originates from AI. A study headed by Valerio Capraro, a psychology professor at the University of Milan-Bicocca, revealed that erroneous AI recommendations not only caused people to become less accurate, but also significantly more confident in their responses (via IBM Think).

      The study also found that individuals who received faulty AI guidance were much less likely to acknowledge their uncertainty, even when the option to remain silent was available, and they stood to gain from being accurate.

      What causes erroneous AI advice to instill such confidence?

      Researchers posed six questions to participants regarding obscure movie facts, intentionally selecting those that the AI model consistently answered incorrectly. This approach allowed them to focus on the effects of consulting AI, irrespective of the quality of its advice.

      The findings were notable. In the absence of AI, participants opted not to answer around 36 to 44% of the time across two experiments. When AI was present, that number plummeted to just 3 to 6%. Accuracy also suffered, decreasing from 27.5% correct without AI involvement to just 9.2% when AI was used.

      Capraro suspects that AI might reduce the level of confidence required before individuals commit to an answer, effectively altering their self-assessment of certainty. Interestingly, this was true even when AI suggestions appeared automatically without being sought, implying that the effect is not merely linked to the act of seeking help.

      Does increasing the stakes alter the situation?

      Introducing financial incentives and penalties led to a slight increase in caution and accuracy, although the underlying trend remained consistent. Capraro clearly differentiates here; he advocates for AI to enhance human judgment rather than replace it entirely.

      He is concerned that this phenomenon might also affect children, as growing up with immediate answers might hinder their experience of doubt, which is an essential component of learning. Recent reports suggest that the younger generation, accustomed to AI, is even relying on chatbots for assistance in face-to-face conversations. As Capraro observes, doubt isn't a sign of lacking knowledge; rather, it often marks the beginning of genuine understanding.

      Manisha Priyadarshini is a technology and entertainment writer with over nine years in editorial experience.

      Mark Zuckerberg envisions AI superintelligence accessible to everyone.

      Meta claims it can provide advanced AI to billions, although users won’t have control over the underlying infrastructure, protections, or restrictions associated with it.

      In an opinion piece for The Wall Street Journal, Mark Zuckerberg outlines a simple goal for Meta's AI: to make superintelligence widely available, equipping everyday individuals with technology that can enhance their health or assist in their work. Zuckerberg proposes personal superintelligence as a counterpoint to sophisticated AI being managed by a select few entities. Meta certainly has the capacity to distribute this on a scale unmatched by most companies, but what users would actually receive is an assistant developed and regulated elsewhere, with crucial decisions made on their behalf.

      The US government has just banned foreign-made robots and power inverters.

      The government has stated it wants to decline the use of foreign robots and power inverters, at least for new models.

      After months of prohibiting all foreign routers and then later approving some of them, the FCC has turned its attention to robots and power inverters. The Federal Communications Commission has recently added two new items to its Covered List, specifically targeting advanced robots and power inverters produced outside the United States. This decision follows a security assessment led by the White House that identified both categories as legitimate threats to national security.

      I tested Dell’s sleek XPS 13, and it provides a refreshing option among budget Windows laptops.

      Dell has managed to reduce the price of the XPS 13 without compromising its premium appeal.

      When Apple introduced the MacBook Neo, it changed my expectations for budget laptops. With a starting price of $599, it combined an aluminum chassis with the quality that one associates with a MacBook, setting a challenging benchmark for Windows manufacturers. Although that price has increased since then, it remains a solid offer for what is currently listed. During Computex 2026, I encountered a potential contender to the MacBook. After spending time with various budget Windows laptops at the event, I noted that the category seemed poised to move beyond plastic designs and subpar displays, and one of the models I was particularly eager to explore was the Dell XPS 13.