What makes AI today so much more advanced than its predecessors?

In the first episode of Mike and Amit’s mini-series on artificial intelligence, our hosts discussed the birth of AlphaGo and the technological underpinnings of generative AI.

But what makes ChatGPT different? Is there anything that fundamentally sets this latest generation of AIs apart from their predecessors?

At heart, says Amit Joshi, generative AI is still based on probability. Based on a given input a machine will try to figure out what should come next, and programs based on neural networks and backpropagation were brilliant at solving problems where context was not important.

But now AI has moved on. It can understand complex ideas without the framework of a board game or a closed system. That, says Mike Wade, is a game-changer.

Authors

Amit Joshi

Professor of AI and Strategy

Amit Joshi is Professor of AI and Strategy at IMD and Program Director of the AI Strategy and Implementation program, Generative AI for Business Sprint, and the Business Analytics for Leaders course. He specializes in helping organizations use artificial intelligence and develop their big data, analytics, and AI capabilities. An award-winning professor and researcher, he has extensive experience of AI and analytics-driven transformations in industries such as banking, fintech, retail, automotive, telecoms, and pharma.