As the tech industry rushes toward a future where the word "AI" is added to every new product release and development, Intel CEO Pat Gelsinger offered an interesting perspective on the future of AI at the World Economic Forum this week. When asked about AI R&D, he referred to Daniel Kahneman's book "Thinking, Fast and Slow" and discussed the benefits of applying the book's basic principles to AI.

Kahneman distinguishes between two ways of thinking, in his view: fast and intuitive and slow and rational. Gelsinger notes that the current state of AI development is that all AI systems today are "fast thinkers" and that incorporating rational thinking and reasoning into AI, or "slow thinkers," is a major area of current research.

"All AI systems today think fast, and we are not bringing reasoning into AI. Today our systems are hallucinating, but tomorrow they must be right if we are to use them extensively."

Given the phenomenal speed with which current AI models can generate results, tracking the authenticity of the data produced may seem like a huge task, although progress has been made in this area. However, creating models that allow the models themselves to bring inference into the equation without having their homework checked may be the next step in the evolution of AI.

Thinking "fast and right," as the Intel chief puts it, will open the door to the adoption of AI that is more useful with respect to the results produced and more reliable with respect to the level of task applied.

AI is currently being implemented in real-world products that require a level of trust in system reasoning, such as autonomous vehicles, with mixed results, sometimes truly catastrophic.

I feel that AI today is becoming more of a buzzword associated with product launches than a truly useful game changer in the applications to which it is currently applied. [As we move forward into the unknown, and as the debate continues about the morality and ethics of our commitment to a technology that we often struggle to define, whether such research and development into the "correctness" of AI-generated results will yield usable results, and far more useful implementations It will be interesting to see.

Only time will tell, but in the meantime, let's hope that the great minds working on the next generation of AI will produce the kind of results that science fiction has long dreamed of, and that there will be fewer AI lapel pins strapped to the front of our shirts.

One can only dream...