Term of the Moment

Ethernet


Look Up Another Term


Definition: AI in a nutshell


Today, AI is primarily used for pattern recognition of faces, speech and trends. It is also employed to make decisions in business and make profits in the financial markets. See big data and high-frequency trading.

The Tools
The architecture of AI software is the neural network, which is unlike regular "if this-do-that" programming. Comprising multiple layers that connect to each other, the neural name was coined as a very imprecise analog of the neurons in the human brain. People learn by experience as do neural networks, which analyze patterns in both known and unknown data. The more layers, the "deeper" the learning. See deep learning and neural network.

The Concern for the Future
What worries people in the industry is the research being done to replace human decision making with AI software. For example, should an AI program be used to determine which products a business should sell or when the company should expand or contract? More significantly, do we leave the decision to go to war up to a machine? There is a huge amount of controversy regarding the future of AI (see technology singularity and AI anxiety). See AI and AGI.




AI Will Reign Supreme
MIT professor Max Tegmark's best-selling book postulates an AI that far exceeds human intelligence and literally takes over.