Roger Schank asserts that the idea of machines thinking and the need to worry, regulate, or grant them civil rights is unfounded. He discusses two predominant approaches to AI: one that mimics human behavior and one that relies on fast statistics-based computing. He argues that neither approach will lead to machines creating a new society.
Schank uses IBM’s Watson as an example, highlighting its limitations in understanding context and emotions. He emphasizes that keyword search and number crunching are not equivalent to genuine thinking or understanding.
He explains that AI researchers who aim to copy human capabilities are working on understanding various aspects of the human mind, such as language processing and learning. He provides an example of a computer program that mimics human memory organization to offer relevant stories but clarifies that the computer doesn’t genuinely understand what it’s saying.
Schank dismisses fears about AI taking over the world and forming robot societies, stating that current AI capabilities are far from achieving such complex tasks. He believes that AI’s primary applications would be as household robots or personal servants, provided they can understand human interactions and avoid annoying behavior, which is beyond the current state of AI.
In conclusion, Roger Schank contends that there is no reason to fear AI, as current AI technology is not advanced enough to pose significant threats. He suggests that people should be more optimistic about AI’s potential benefits.