Understanding AI’s Impact: Balancing enthusiasm, skepticism, and positive social change
AI is not a substitute for human intelligence, creativity, or judgment. It can improve productivity and decision-making, but only when used carefully. Like any powerful technology, its value depends on how well people understand its strengths, limits, and risks.
Case in point: Google AI Overviews save time by synthesising complex information from multiple web sources into a single, comprehensive snapshot at the top of search results. They enable faster multi-part research, provide direct inline citations for easy factchecking, and handle nuanced conversational queries effortlessly. At the same time, these overviews can put people at risk of harm with inaccurate and misleading information and advice. Although 91% of overviews had correct information, only of it was both correct and supported by cited sources, a combination that’s deemed “trustworthy”.
This is why understanding AI applications and systems is so important. It is easy to focus on what these systems can do, but positive impact depends on knowing how they work, where they are useful, and where they fall short. A system may appear confident and still miss important context. It may produce useful predictions while overlooking the people most at risk.
Healthy skepticism
Healthy skepticism plays an important role in one’s outlook towards AI. Skepticism should not be seen as resistance to innovation. It is a necessary part of responsible use. It asks important questions: Who benefits from this system? Who might be left out? What assumptions is it making? What happens if it gets things wrong? These are not abstract concerns.
They affect safety and public trust.
AI systems often work well in some contexts and fail in others, which means the public is asked to judge risk without always having full visibility into how the system arrives at its outputs. In that setting, trust becomes less about blind confidence and more about whether people believe the system is reliable, well-governed, and honest about its limits.
AI comes with uncertainty
Public understanding of risk is equally important because it shapes how people respond to AI in everyday life. If risks are framed too narrowly, people may overlook possible harms; if they are framed too alarmingly, they may dismiss useful applications altogether. An optimal risk perspective therefore depends on helping the public understand that uncertainty is not a flaw to be hidden, but a feature to be communicated clearly. When people are given the tools to interpret uncertainty, question outputs, and recognise where human judgment is still needed, they are better able to build informed trust rather than passive dependence.
In the end, the value of AI lies not in how advanced it appears, but in how responsibly it is used. When AI systems are understood well, questioned carefully, and kept grounded in human needs, they can support and lead to real progress.
As Sam Altman, CEO, OpenAI said, “AI won’t replace humans, but humans who use AI will replace those who don’t.”
Upcoming AI and Society Speaker Series
The LRF Institute for the Public Understanding of Risk (IPUR) and Institute of Behavioral Science and Technology (IBST) have co-launched the AI and Society Speaker Series series looking at not only on what AI can do but also on how people perceive, interpret, and respond to the risks it creates. The series brings together leading scholars, practitioners, and policy thinkers to examine how AI is reshaping risk, trust, and decision making across society. IBST’s research sits right at the intersection between behavioral science and emerging technology.
The next webinar is on 13 October and will be presented by Prof Carl Benedikt Frey, Dieter Schwarz Foundation Professor of Political Economy and Technology at the Oxford Internet Institute at the University of Oxford.
