Beyond the Hype – Why Today’s AI Still Can’t Think Like a Human—and Why It Matters
Artificial Intelligence (AI) has taken enormous strides in recent years, especially with the rise of large language models (LLMs) such as ChatGPT and similar generative systems. These AI systems excel at producing human-like text, answering questions, and assisting with creative tasks by predicting the most probable next word or phrase based on vast amounts of data. However, despite these impressive capabilities, today’s AI fundamentally differs from human intelligence in crucial ways.
Five Fundamental Human Brain Skills AI Has Not Replicated
Current AI systems, focused largely on improving predictive text models, lack essential cognitive abilities that humans naturally possess. These skills are critical to how humans understand and interact with the world but remain beyond the reach of AI built on statistical pattern recognition:
- Causal Reasoning and Understanding
Humans instinctively grasp cause and effect. We form mental models that allow us to hypothesise about unseen events, anticipate consequences, and understand why things happen, not just what happens next. In contrast, AI systems rely on correlational data without genuine understanding of causality, limiting their ability to reason about novel scenarios or counterfactual conditions. - Common Sense and Contextual Awareness
Everyday human reasoning draws from a rich, implicit knowledge of how the world works—physical laws, social norms, and practical experience. This common sense enables us to fill gaps, make inferences, and interpret vague or ambiguous information. AI models, however, struggle when context is incomplete or unstated, often producing outputs that defy practical sense or real-world logic. - Mental Simulation and Imagination
Humans can imagine hypothetical futures, simulate actions mentally, and creatively explore possibilities. This capacity enables planning, problem-solving, and innovation. AI tends to generate responses based on learned patterns rather than actively imagining or simulating new ideas beyond its training data. - Active Learning and Development
Human intelligence develops continuously through active interaction with the environment from birth onward, learning adaptively from experiences. In contrast, AI training is generally static—models are trained offline on large datasets and deployed without ongoing adaptive learning. They lack the kind of experiential development that shapes how humans grow smarter and more nuanced over time. - Flexible Generalisation and Knowledge Transfer
Humans apply learned knowledge flexibly across domains, quickly adapting skills learned in one context to solve problems in another. AI systems typically excel only within the scope of their training data distributions and often falter when faced with unfamiliar or out-of-distribution situations.
Taken together, these five foundational skills highlight the gap between today’s predictive AI and the richer, more versatile intelligence of the human brain.
The Intensifying USA-China AI Race – What’s at Stake?
Behind the headlines about AI breakthroughs is a fast-growing rivalry between the United States and China. This “AI race” is framed as a high-stakes competition for technological supremacy, economic power, and military advantage. Each country invests heavily in AI research, infrastructure, talent acquisition, and industrial deployment.
The US continues to lead in fundamental AI research and innovation, supported by a vibrant private sector and academic institutions. China is rapidly catching up, harnessing centralised government coordination to push AI applications, patent filings, and nationwide infrastructure development. Both nations view AI as critical to future competitiveness and influence.
However, this race narrative can oversimplify and even obscure more pressing challenges:
- Overshadowing Deeper AI Challenges:
The rush to dominate AI technology often focuses on speed, hardware scale, and narrow application wins while neglecting the deeper cognitive challenges—like enabling machines to truly understand cause, context, and imagination—that must be solved for AI’s transformative potential. - Risk of Short-Term, Zero-Sum Thinking:
Framing AI development as a zero-sum game risks prioritising quick victories over responsible innovation. It can encourage a “race to the bottom” mentality where ethical, societal, and long-term risks are sidelined. - Leadership and Governance Gaps:
Selling the AI race as an unequivocal good can mask a lack of nuanced and “smart” leadership. What’s needed is balanced governance that fosters competitiveness without compromising ethical standards or global cooperation on safety, fairness, and transparency. - Missed Opportunity for Collaboration:
Many of the challenges AI poses—such as bias, misinformation, security, and respect for human values—are global in nature. Constructive international collaboration could help address these issues more effectively than adversarial competition alone.
Towards Smarter Leadership and AI Beyond Prediction
Current AI progress is remarkable but incomplete. Large language models and similar systems, fundamentally predictive in nature, have not yet demonstrated the uniquely human skills of causal understanding, common sense, imagination, active learning, and flexible generalisation.
Meanwhile, the growing USA-China AI race brings real investments and innovation but also risks reinforcing a shortsighted focus on dominance rather than meaningful, responsible AI breakthroughs. Rather than celebrating speed and scale alone, the true promise of AI lies in moving beyond purely statistical prediction to cognitive architectures that embody the deeper mental abilities humans use every day.
For policymakers, researchers, and the public, this means advocating for smarter leadership—guiding AI development with vision, ethics, and collaboration—not simply a contest mentality. Only by addressing these fundamental gaps can AI fulfill its potential as a transformative technology that truly augments human intelligence.
The flow on from this current narrow focus of predictive AI development, at all costs, is that many governments are pushing for AI to be introduced into schools and the workplace. This is destined to crush human development permanently, as people forget how to really “think like a human”.