In the 1980s, the following researchers first articulated and discussed the AI paradox.

• Hans Moravec - Robotics Institute of Carnegie Mellon University in Pittsburgh, USA,  
• Rodney Brooks - Massachusetts Institute of Technology, and  
• Marvin Minsky - Massachusetts Institute of Technology's AI laboratory

In essence, AI can quickly learn to perform tasks considered "difficult" by humans, such as complex statistics and analysis. However, on the other hand, tasks that are trivial for humans, such as colour recognition or facial recognition, can be exceedingly difficult for computers. In short, computers, algorithms, and AI are excellent at some tasks, but humans excel at others. Therefore, how can we strike the proper balance?

What is Moravec's paradox?

Moravec's paradox is a phenomenon that occurs when we use AI-powered tools. It observes that humans find complex tasks simple for AI to learn. That is when compared to simple sensorimotor skills that humans possess naturally.

AI, for example, can solve complex logical issues and perform advanced mathematics. However, the 'simple' skills and abilities we gain as babies and toddlers - perception, communication, movement, and so on — require significantly more computing to be replicated by an AI.

In other words, for AI, the difficult is simple, and the simple is difficult.

What is the logic behind it?

But why does AI have such a hard time with the simple? Researchers can explain Moravec's dilemma in terms of evolution, comprehension, and perception. To begin with, the abilities we consider "easy" — those we learn spontaneously — result from thousands of years of evolution. 

Put another way, the intricacy of the basic abilities we take for granted is hidden. Furthermore, AI 'learns' by listening to our instructions on doing things. We've consciously learned how to solve problems, win games, and apply logic. The steps (computations) required to execute these jobs are known. As a result, we can teach AI to them.

Moravec's paradox and AI in the past

Moravec's paradox has had an impact on AI throughout its history. It may be one factor that has slowed development and led to the AI effect.

The AI effect is a phenomenon in which AI-powered technologies lose their 'AI' title over time since they aren't brilliant. Moravec's paradox may have played a role in this. The activities these gadgets do are easy once broken down, so they lose their 'intelligent' designation. Because of Moravec's paradox, no matter how professional AI tools and programs were at games and reasoning, they couldn't do 'basic' human activities.

How can anything be 'intelligent' if it can't mimic a toddler's behaviour?

Moravec's paradox and current AI

Moravec's paradox explains why AI that can reason at an adult level is an old hat, but AI that can see, listen, and learn is new and exciting. For example, image classification and facial recognition — the machine's version of sight — are becoming more common.

Meanwhile, personal assistants such as Alexa are examples of AI's ability to 'hear' and understand us. As with these assistants or breakthroughs like Google Duplex, AI is increasingly capable of speech.

Conclusion

According to Moore's law, computers were hundreds of millions of times quicker in the 2020s than in the 1970s. The extra computer power was finally enough to handle perception and sensory skills, as Moravec predicted in 1976. 

Likewise, Brooks decided to explore a new route in artificial intelligence and robotics research in the 1980s as a result of this. He decided to create intelligent machines with "no cognition." Instead, it's only a matter of perceiving and acting. Similarly, successful AI applications in the twenty-first century do not replicate step-by-step "intelligent" problem resolution. Instead, the machines mimic people's quick, "intuitive" decisions that allow them to spot patterns and anomalies instantaneously and automatically.

Moreover, AI has experienced its share of highs and lows. It is a field rife with ethical dilemmas, difficulties, and the occasional paradox.