From the course: Introduction to Artificial Intelligence

Strong vs. weak AI

From the course: Introduction to Artificial Intelligence

Strong vs. weak AI

“

- So, when is a computer system intelligent? We've seen that when a computer system is just matching symbols, it's almost like a high-tech phrase book. The system might seem intelligent, but it's actually just like a parrot with a great memory. The philosopher John Searle said that you can think of artificial intelligence in two ways. There's strong AI and weak AI. He thought we were much further away from intelligent systems than most people realize. Strong AI is when a machine displays all the behavior you'd expect from a full fledged person. This is usually what you see in science fiction. These are artificial beings that have emotions, a sense of humor, and even have a sense of purpose. That's why C-3PO is scared when they land on Tatooine. It's also why Commander Data shows real creativity when battling the Romulus. On the flip side, there's weak AI. A personal assistant like Apple Siri is a good example of weak AI. This is AI that's confined to a very narrow task. It's like when a system processes language into text or when it sorts all the pictures on your computer. Most AI experts believe that we're just starting down the path of weak AI. Think about Siri. You can talk to Siri and ask questions. Siri listens to your input and then converts your language into something that the computer recognizes. Then, Siri matches a response to what's in her database. Most of the energy right now in AI is around developing and expanding weak AI. Strong AI is still very much just science fiction. In the 1970s and '80s, symbolic systems were used to create weak artificial intelligence. These were commonly called expert systems. In these systems, you have experts create a list of steps to solve a complex problem. If the list is long enough, it starts to seem like intelligence. But again, this system is just parroting back program responses created by experts. These expert systems were often used in medicine. A nurse might input symptoms into a computer. If the patient has a cough, then check if they have a temperature. If they have a cough and a temperature, then check to see if they're dehydrated. If they have a cough, a temperature, and are dehydrated, then ask the nurse to check for bronchitis. To a patient, it might look like they're being diagnosed by an intelligent computer. In reality, the program is just matching the symbols and patterns that an expert created to reach a diagnosis. Just like the phrase book in the Chinese Room experiment. In the end, expert systems ran into the same problem as any other symbolic system. They would lead to explosions of combinations. There were just too many patterns to match. Think about all the different questions a doctor might ask to reach a diagnosis. Yet, the symbolic systems approach was a key starting point for artificial intelligence. It's still used today. In fact, many experts still refer to it as GOFAI, or good old-fashioned AI.

Contents