In the last few years, we've seen the rise of AI that is more focused on people, like Alexa, Siri, and Tesla's autopilot. Businesses, too, are getting into AI. Machine learning, a type of narrow AI, is on the rise. It searches through a lot of data to make it easier to make and market products. Additionally, computers have also won games that were very difficult in the past, like alphaGo.

What is synthetic intelligence?

In terms of philosophy, synthetic intelligence (SI) is a better definition of what AI (which encompasses machine learning, deep learning, and reinforcement learning) entails. Artificial indicates that something isn't "real," however ML, DL, and RL are real since they exist in the actual world, have a variety of real-world effects, and do a variety of real-world tasks.

John Haugeland compares simulated and synthetic diamonds, claiming that only the synthetic diamond is genuinely a diamond. Synthetic refers to something created by synthesis, the process of joining pieces to make a whole; informally, a human-made replica of something that has evolved naturally. Thus, a "synthetic intelligence" would be or appear to be created by humans but would not be a simulation.

Haugeland coined the word in 1986 to refer to prior AI research, which he dubbed "good old fashioned artificial intelligence" (GOFAI). The pioneers of AI were adamant that their techniques would result in computers with accurate, human-like intelligence. Following the first AI winter, many researchers focused away from artificial general intelligence and on solving specific individual issues, such as machine learning, an approach dubbed "weak AI" or "applied AI" by specific popular sources. The term "synthetic AI" is now occasionally used by researchers in the field to distinguish their work. These employ sub symbolism, emergence, Psi-Theory, and other relatively new methods for defining and creating "true" intelligence from previous attempts, most notably those involving GOFAI or weak AI.

Researcher's opinion

Sources disagree about what "real" intelligence is and whether there is a meaningful difference between AI and synthetic intelligence.

• Russell and Norvig specify the following example:

1. "Are machines capable of flight?" Because aeroplanes fly, the answer is yes.
2. "Are machines capable of swimming?" Submarines do not swim. Hence the answer is no.
3. "Are machines capable of reasoning?" Is this question similar to the first or identical to the second?

• Drew McDermott feels that we should interpret "thinking" similarly to "flying." He contends, "Saying Deep Blue doesn't truly think about chess is like saying an aeroplane doesn't fly because it doesn't flap its wings," when discussing the computer chess champion Deep Blue.
• "Whether machines can think is as relevant as to whether submarines can swim," Edsger Dijkstra agrees.
• "No one thinks computer simulations of a five-alarm fire will destroy the neighbourhood or that a computer simulation of a rainstorm will leave us all drenched," says John Searle. It's one of the main points of his argument that a natural person is very different from a computerized person.
• Daniel Dennett feels that this is essentially a semantic disagreement unrelated to the critical issues in AI philosophy.

Conclusion

SI is a more relevant category for machine learning, deep learning, and reinforcement learning because synthetic denotes anything generated by intelligent agency rather than undirected natural processes.