The Illusion of Machine Thinking
How exactly do we measure a machine’s ability to think? Recent advancements in AI have sparked debate about whether machines can truly replicate human cognition. Large language models demonstrate impressive capabilities in processing vast amounts of data, but they’re missing something essential—the ability to make creative mental connections that come naturally to humans.
The numbers are eye-opening. About 72% of businesses now use AI in some form, with economists predicting massive economic impacts by 2030. Recent surveys show that 78% of organizations reported AI usage in 2024, up from 55% the previous year. Everyone’s jumping on the AI bandwagon. But here’s the kicker: these systems aren’t actually “thinking” like we do.
AI models struggle embarrassingly with analogical reasoning. Put a story-based problem in front of them, and they falter. Humans—even those without specialized training—consistently outperform AI in tasks requiring creative analogies and zero-shot learning. No contest. The machines simply can’t compete when it comes to making mental leaps without pre-existing data.
AI’s Achilles heel: creative leaps and analogical reasoning remain uniquely human domains where machines fall flat.
Another fascinating weakness? AI is weirdly susceptible to answer-order effects. Change the sequence of options, and you might get entirely different results. Consistency? Not their strong suit. Historical biases in training data continue to produce flawed decision-making outcomes across critical domains.
The tech experts are worried, too. By 2035, they predict AI will negatively impact fundamental human traits like empathy, moral judgment, and social intelligence. Great. Just what we need—technology that makes us less human.
Despite all the hype about artificial general intelligence, current AI systems are glorified parroting machines. They repackage existing information rather than generate truly original thought. They lack metacognition—they don’t know what they don’t know. Kind of a big deal when you’re claiming machines can “think.”
Some optimists believe human-AI collaboration might lead to new forms of creativity and problem-solving. Maybe. But reliance on these systems raises serious questions about human autonomy. Do we really want to outsource our thinking?
The so-called “2025 AI Index Report” highlights advances relevant to policymakers, while corporations increasingly recognize AI’s role in driving growth. According to research published in February 2025, AI models consistently show poor performance in letter-string and matrix analogy problems compared to humans. But at what cost? The pursuit of “Authentic Intelligence” attempts to address ethical gaps, but we’re nowhere close to solving them.
Let’s be clear: AI doesn’t possess the emotional intelligence, social awareness, or capacity for deep thinking that defines human cognition. It’s a tool—impressive but fundamentally limited. Can AI think like us? Not even close. And maybe that’s okay.
