Artificial Intelligence (AI) is a field that has grown tremendously in the last decade, and there are many discussions and fears about how AI may replace humans in nearly any job. While we agree that AI can help humans in many jobs, and increase automation in several fields, humans are far to be replaced, as Artificial Intelligence is not as Intelligent as you are often told.
AI does not really function like the human brain
Deep learning systems are not designed to imitate the human brain. While the human brain is incredibly complex, deep learning systems are designed to learn from data and make predictions based on that data. For example, a child does not need to be shown hundreds or thousands of cats before understanding what a cat is, as deep learning systems need. This is because humans have a natural ability to learn and understand concepts, while deep learning systems require vast amounts of data to learn statistical patterns.
A scientific demonstration
A scientific paper (Gleave, Adam, et al. “Adversarial policies: Attacking deep reinforcement learning.” (2021)) is an interestin example of how AI does not understand “the big picture.”
In a game of Go, two players alternately place black and white stones on a board marked out with a 19×19 grid, seeking to encircle their opponent’s stones and enclose the largest amount of space. The huge number of combinations means it is impossible for a computer to assess all potential future moves.
AlphaGo, a system devised by Google-owned research company DeepMind, defeated the world Go champion Lee Sedol by four games to one in 2016, leading the champion to retire, as he believed that AI systems would be unbeatable for humans. He was wrong.
Recently, the system was easily beaten by an amateurish strategy, that would unlikely work against a human. The tactics used involved slowly stringing together a large “loop” of stones to encircle one of his opponent’s groups, while distracting the AI with moves in other corners of the board. The Go-playing bot did not notice its vulnerability, even when the encirclement was nearly complete.
This example highlights the fact that AI does not understand the context or the big picture. While AlphaGo was designed to play the game of Go and had become proficient at it, it was unable to adapt to new strategies or tactics that were not part of its training data. This is because deep learning systems can only make predictions based on the data they have been trained on and cannot make decisions based on context or intuition.
Another example that demonstrates the limitations of AI is self-driving cars. While self-driving cars have come a long way in recent years, they still require human intervention in certain situations. This is because AI does not have the ability to understand the nuances of driving or the ability to make intuitive decisions based on context. While AI can recognize objects such as other cars, pedestrians, and traffic lights, it cannot understand the context of the situation, such as a child running after a ball or a car suddenly changing lanes.
While AI has come a long way in recent years, it still has limitations that must be taken into consideration when designing and implementing AI systems. As AI continues to evolve, it is essential to understand its limitations and work towards developing new technologies that can overcome these limitations.