Thank you! This is a great and very interesting article. Two questions come up for me. First—Could causal AI be used in replace of machine learning systems that are making predictions and used in the public sector (e.g., related to predictive policing)? Or would it more be used to inform interventions to address root causes rather than serve as decision making support tool?
Second, it seems that similar to AI that uses machine learning, one of the issues is that the quality of the results relies on the quality of the data put in. Is that right?
This is an excellent summary of an especially important aspect of AI research. While "deep learning" has become incredibly useful and common since the advent of sufficiently powerful computers—the fundamentals of it are not new at all, but decades-old—it is neither "learning" nor "deep" in any human sense. A much better phrase would be "mass automation of object perception," the ability to train computers to recognize the same kinds of entities and patterns as humans. Causal AI, in contrast, strikes deep at the roots of what distinguishes humans from other forms of life: The ability to perceive how one action creates another, and thereby learn how to manipulate the world around them to their advantage. Unfortunately, even the theory of how to do this remains quite primitive. The two techniques mentioned are, for example, not much more than ways to describe causality with more precision once it has been uncovered… and that, alas, still has to be done mostly by humans, and so is not easily automatable.
Artificial Intelligence is building an virtual world for all the products and services creating a reality experience of it. AI is the controlled computer operation or robot that fulfills the need of completing a particular task in quick time. AI has become the trend as well as need of every individual as well as businesses and as a result companies are continuously working on AI to develop remarkable solutions for the market.
COMMENTS
BY Genevieve Smith
ON June 5, 2020 01:49 PM
Thank you! This is a great and very interesting article. Two questions come up for me. First—Could causal AI be used in replace of machine learning systems that are making predictions and used in the public sector (e.g., related to predictive policing)? Or would it more be used to inform interventions to address root causes rather than serve as decision making support tool?
Second, it seems that similar to AI that uses machine learning, one of the issues is that the quality of the results relies on the quality of the data put in. Is that right?
BY Terry Bollinger
ON May 26, 2021 07:17 PM
This is an excellent summary of an especially important aspect of AI research. While "deep learning" has become incredibly useful and common since the advent of sufficiently powerful computers—the fundamentals of it are not new at all, but decades-old—it is neither "learning" nor "deep" in any human sense. A much better phrase would be "mass automation of object perception," the ability to train computers to recognize the same kinds of entities and patterns as humans. Causal AI, in contrast, strikes deep at the roots of what distinguishes humans from other forms of life: The ability to perceive how one action creates another, and thereby learn how to manipulate the world around them to their advantage. Unfortunately, even the theory of how to do this remains quite primitive. The two techniques mentioned are, for example, not much more than ways to describe causality with more precision once it has been uncovered… and that, alas, still has to be done mostly by humans, and so is not easily automatable.
BY Amit Pawar
ON January 24, 2022 08:43 PM
Artificial Intelligence is building an virtual world for all the products and services creating a reality experience of it. AI is the controlled computer operation or robot that fulfills the need of completing a particular task in quick time. AI has become the trend as well as need of every individual as well as businesses and as a result companies are continuously working on AI to develop remarkable solutions for the market.