Artificial General Intelligence/AGI

General artificial intelligence refers to artificial intelligence that has the same intelligence as humans or surpasses humans, and can exhibit all the intelligent behaviors of normal humans. It is one of the main goals of artificial intelligence research and a major topic discussed by science fiction novels and futurists.

Operational means of testing general artificial intelligence

What kind of testing standards does general artificial intelligence need to pass? Currently, there are mainly the following testing schemes:

• Turing test: proposed by Turing, mainly used for communication with humans;
• Coffee test: Proposed by Goetzel, it is mainly used to test space and operational skills in life. It relies on the calculation of robotics and image recognition. It brings the machine to ordinary American families and lets it know how to make a cup of coffee without deliberate design. It needs to actively recognize the coffee machine in an unfamiliar space, identify coffee and water, find the right cup and put it away, and then press the correct keys and operations to brew coffee.
• Robot Student Test: Proposed by Goetzel, it is mainly used to analyze and answer single-question tests through machine learning;
• Employee test: Proposed by Nielsen, it is used to test the ability to coordinate, infer, imagine, and plan to solve complex problems.

These tests will measure a range of necessary traits, including reasoning and learning abilities.

Problems that general artificial intelligence needs to solve

The most difficult problems for computers are called AI-Complete or AI-Hard.

Therefore, solving these computational problems is equivalent to solving the core problem of AI - making computers as smart as humans or general artificial intelligence.