We continue to debunk the myths about the future of the human activity. Today – the myth of Intelligence.
Until recently, it was thought that driving a car, diagnosing and recognizing birds could not be automated. However, today there are cars with autopilot, systems for diagnosing, and even an application for instant identification of birds.
The reason for such erroneous judgments was the belief that machines should copy people’s way of thinking. And that the only way to automate tasks is to ask a specialist to explain in detail and step by step how he does his job, and then try to record these actions in the form of instructions. Let’s analyze this on the example of diagnosis.
If you ask about it, you will hear some basic rules, but in general, they will be difficult to explain. Diagnosis requires qualities such as creativity, judgment, and intuition – things that are difficult to formulate in the form of step-by-step instructions. And if a person can’t explain exactly how he does his job, how can he explain it to the car? 30 years ago such an opinion was correct, today it is doubtful, but in the future, it will be completely wrong.
How can an algorithm diagnose?
A few years ago, a team of researchers from Stanford developed a system that can determine with very high accuracy whether freckles are cancerous. This works as follows: the system uses a pattern recognition algorithm, comparing a specific case with about 130 thousand previous cases, and looks for similarities between them. The program does not try to copy the doctor’s judgments or thinking, and does not understand medicine at all. However, the algorithm is able to analyze more cases than any doctor could in his lifetime.
The myth of Intelligence shows that our idea of the human mind, of how we think and make decisions, is not a limitation for automation. Human ability is not the pinnacle of what machines can do in the future.