When AI becomes your default adviser, what happens to your judgment?

AI now sits inside everyday decisions: what to buy, where to go, how to learn, and sometimes how to cope. Executives tend to focus on productivity. The harder question is what happens when convenience turns into deference, and an assistant becomes the authority behind a decision.

Almost every day, new research and news stories serve as cautionary tales: In China, a farmer lost 25 acres of seedlings after asking an AI for advice. Teens are turning to AI chatbots for companionship and mental health advice. People are turning to AI for babysitting. Hikers were stranded atop a mountain after AI-generated instructions left them improperly equipped. People fell ill after eating poisonous mushrooms that were misidentified by AI.

The examples do not prove that everyday AI use is inherently harmful. They do show why casual assistance must be distinguished from decisions involving safety, health, money, parenting, or emotional vulnerability. In those situations, a plausible answer is not the same thing as a reliable one.

WEAKENING CRITICAL THINKING 

“The growing reliance on AI for emotional support, entertainment, and everyday decision-making raises serious concerns about the quality of our human agency and social relationships,” says Dr. Mohammad Ayish, Professor and Interim Head of the Department of Media Communication at the American University of Sharjah.

And when we routinely outsource reasoning, Samira Cutts, Cognitive Neuroscientist and Head of Brain Health at Prime Performance Labs, says, “We risk weakening our capacity for critical thinking and independent judgment.” 

Recent research suggests that using AI too much can harm our ability to think and judge for ourselves, leading to “epistemic de-skilling” and weaker critical thinking. Concerns about how AI affects children’s development even led New York City to ban generative AI in schools.

Meanwhile, the overuse has led to concerns about AI replacing human relationships. While AI can offer companionship and support, Cutts adds that it cannot replicate the reciprocity, emotional complexity, and repair that make human relationships meaningful. 

She says the risks are even greater when it comes to mental health advice. “AI can sound authoritative and empathetic without possessing clinical judgment, genuine understanding, or professional accountability. In mental health, plausible-sounding advice is not necessarily safe advice. AI may offer a useful starting point, but it should never be mistaken for a substitute for qualified professional assessment and care.”

AI also raises bigger questions about how we make decisions, says Prashant Malaviya, Professor of Marketing and Vice Dean of Programs at Georgetown University’s McDonough School of Business. 

He says there are three reasons why people might trust AI’s answers so easily. First, it’s easier than thinking for ourselves. Second, AI’s smooth language and apparent expertise can make its answers seem more trustworthy than they are. Third, we might not know enough to question what AI tells us. 

“The concern is that repeated reliance could weaken the habit of forming and testing our own judgments. AI can inform a decision, but people still need to interrogate its output and retain responsibility for the final judgment,” adds Malaviya.

PROTECT THE CORE HUMAN SKILLS

Even as some question its benefits and tire of its ubiquity, the spread of the technology is inevitable.

Dr. Ayish says AI itself isn’t the problem. Most concerns come up, he says, “when technological convenience is used to replace human judgment, critical reflection, and meaningful interpersonal engagement.”

How AI can change how we think, decide, and interact with the world is well-documented, says Martin J Kraemer, CISO Advisor at KnowBe4. “Cognitive offloading and blind trust have well-documented long-term effects that lead to disengagement and reduced mental flexibility.”

The technology can assist human decision-making without replacing the expertise behind it, like doctors assessing automated diagnoses and pilots using assistance systems while retaining essential flying skills, but Kraemer says the defining challenge of our time is to identify and protect the core human skills. 

“Identifying and protecting the core skills and competencies that we, individually and as a society, desperately need to thrive is one of the most important challenges of our time,” adds Kraemer.

LIMIT THE USE OF AI

In the region, Saudi Arabia and the UAE have embedded AI literacy into their national strategies to drive economic growth, workforce development, and responsible AI adoption. However, as in many countries, people still struggle to understand how the technology works and its limits. 

The enthusiasm is understandable. But as people embrace the technology for everyday guidance, including mental health, babysitting, and even farming advice, should we be thinking about how to guide its use responsibly?

For example, in China, the government is setting rules for AI apps and chatbots to prevent people from relying too much on the technology. It is also introducing new laws to control how AI is used and to stop misuse.

“Recent studies show that consumers often see these tools as confidential advisers,” says Kellie Blyth, Data and Technology Partner at Addleshaw Goddard. However, unlike doctors, lawyers, or financial professionals, these tools do not offer the same legal, regulatory, or professional protections, she adds.

If that information is misused, inadequately protected, or exposed through a cyber incident, Blyth says, “the consequences for identifiable individuals could be severe.”

Regulations should make sure users get clear information about how their data is collected, used, stored, and shared. This helps people decide what they want to share. Providers should also be held accountable and offer solutions if problems occur, she adds.

“While it may be too late to control whether people use AI, it is not too late to establish safeguards governing its use and which aim to protect consumers from the most significant harm.

REGULATION NOT ENOUGH

But regulation alone may not solve overreliance. Malaviya explains that in his research on human advisers, disclosures sometimes made people more persuaded when they saw them as a sign of expertise.

“Whether the same effect applies to an AI ‘expert’ remains an open research question. My hypothesis is that safeguards need to do more than warn people. They should encourage users to pause, question recommendations, and consider alternatives, particularly where the consequences are significant.”

He adds, “We still need research to understand whether overreliance stems primarily from convenience, misplaced trust in AI’s apparent expertise, or users lacking the knowledge needed to challenge its output.”

That points to a more practical standard for responsible AI: design systems that reveal uncertainty, prompt users to verify high-stakes recommendations, and make escalation to a qualified human easy. 

AI can be useful for information and support, Dr. Ayish says, but literacy, ethical awareness and independent thinking remain essential. “We should ensure that AI enhances human capabilities rather than gradually undermining them.”