What Can’t AI Do (Yet)? 20 Examples

Artificial intelligence has become an unavoidable topic of conversation. You’ve probably discussed it with family during the holidays, haven’t you?

This article written by our team aims to explore the aspects that AI cannot master, both on a technical level and on a human level.

20 Things AI Still Can’t Do

Whether through chatbots like our French ChatGPT, recommendation systems, or facial recognition technologies, AI is transforming our lives. But with these advances, it’s easy to fall into the trap of believing that AI can do everything…

However, understanding AI’s limitations is just as important as knowing its capabilities.

1. 1. Understanding cultural context: AI can process data, but it doesn’t understand cultural nuances and specific references that can influence the meaning of a message.
   2. Experiencing authentic emotions: While AI can simulate emotions, it cannot actually feel or understand human emotions like joy, sadness, or love.
   3. Demonstrating true creativity: AI can generate artwork or music by imitating existing styles, but it cannot create something truly original with intentions or emotions behind it.
   4. Making complex ethical decisions: Algorithms can follow rules, but they cannot navigate moral dilemmas where there is no clear answer.
   5. Understanding irony or sarcasm: Despite advances in natural language processing, AI still struggles to detect the subtleties of irony and sarcasm in speech.
   6. Learning autonomously: AI requires datasets to learn. It cannot learn new skills or concepts autonomously without data or instructions.
   7. Accounting for unforeseen variables: AI models are often based on historical data. They can struggle to adapt to unknown situations or unexpected events.
   8. Interacting in a human way: While AI can simulate conversations, it cannot establish genuine emotional connections with people the way a human would.
   9. Understanding complex human emotions: AI can identify emotions in facial expressions or vocal tones, but it cannot truly understand complex emotions like jealousy or nostalgia.
   10. Performing introspection: AI cannot reflect on itself, on its own motivations, or on its existence in a philosophical way.
   11. Handling unstructured tasks: AI excels at structured tasks with clear data, but it struggles with unstructured and ambiguous tasks.
   12. Evaluating beauty or aesthetics: AI can analyze patterns, but it cannot appreciate the beauty of a work of art or the aesthetic feeling of a musical composition the way a human can.
   13. Having common sense: AI lacks common sense and can make obvious mistakes in situations where human intuition would be necessary.
   14. Establishing subjective priorities: AI can rely on quantifiable criteria, but it cannot evaluate priorities that depend on subjective values or personal preferences.
   15. Showing real empathy: While AI can be programmed to recognize and respond to emotions, it cannot experience authentic empathy.
   16. Managing complex social interactions: Human interactions are often loaded with subtle signals and dynamics that escape AI’s understanding.
   17. Making predictions about an uncertain future: AI can analyze past trends to make predictions, but it cannot predict the future with certainty, especially in unpredictable situations.
   18. Applying knowledge flexibly: AI struggles to transfer knowledge acquired from one domain to another, which is something humans do naturally.
   19. Understanding metaphors: AI has difficulty interpreting metaphors and analogies, which require contextual and imaginative understanding.
   20. Interacting with the physical world: While some AI is integrated into robots, the ability to interact with the physical world in a fluid and adaptive way remains a complex and limited task.

See also: Can ChatGPT Replace a Psychologist?

What AI Will Never Be Able to Do

While artificial intelligence has made significant progress in many areas, certain capabilities will always remain inaccessible to it. AI, however advanced it may be, will never be able to feel a true “lived experience,” such as family or health problems.

Even if we could program responses simulating emotions or create scenarios where a bug is interpreted as an illness, the lived experience will remain fundamentally different.

By identifying these boundaries, we can better appreciate the unique role of human intelligence and the challenges to overcome for AI to truly enrich our lives!

Also read: Is ChatGPT Reliable? Here Are the 5 Points to Watch

And in the Distant Future, Will AI Be Able to Do Everything?

Artificial intelligence continues to transform our world, but it has well-defined limitations that should not be underestimated. The 20 examples we examined highlight not only the technical challenges AI must face, but also the crucial importance of human intelligence in areas such as empathy, creativity, and ethics.

Will AI one day manage to evolve to overcome these obstacles?

We can envision that AI will continue to progress and become more “human.” However, even if personalities like Elon Musk manage to create robots with a human appearance, these entities will technically remain machines.

While we can imagine a robot with feelings that aspires to rebel, like in a Terminator-style movie scenario, it’s likely that this will only materialize in several years. What do you think? Share your opinion in the comments!