Many leaders might be tempted to dismiss ChatGPT as one of those gimmicks of new technology — a sort of robot that can by turns impress with its ability to apparently hold a “normal” conversation with the person asking it questions and infuriate when it gets caught out. But this would be a mistake. ChatGPT is significant because it is an example of Artificial Intelligence, the technology that is still creeping up on society but which, according to proponents, has the potential to change things like little that has come before. As long ago as 2018, Google CEO Sundar Pichai was quoted as saying: “AI is probably the most important thing humanity has ever worked on. I think of it as something more profound than electricity.”
That remark appears in Power and Prediction, the latest book from the team that produced Prediction Machines, which was subtitled “The Simple Economics of Artificial Intelligence.” This appears curious because the titles of the books suggest a downplaying of AI’s importance. After all, if this technology is just in the business of prediction then surely it is not so different from forecasting — and we all know how good that is.
The authors — Away Agrawal, Joshua Gans and Avi Goldfarb of the Roman School of Management at the University of Toronto — acknowledge that not everybody has been struck by the effect of AI. Many have worried about the effect of automation on jobs — although, to be fair, less is heard about that now that advanced economies like the U.S. and the U.K. are experiencing worker shortages — while others have pointed out that AI has not yet produced the expected productivity benefits. The latter, they insist, is a not unprecedented paradox. It echoes the situation in the late 1980s when computers seemed to be everywhere without any measured improvements in productivity. It was also the case when steam engines and electricity first appeared. What they call general purpose technologies can seem to be slow to take off and then end up transforming economies, businesses and work itself. We are, argue Agrawal and his colleagues, in “The Between Times.” Some businesses are already seeing financial benefits in being able to carry out enhanced prediction. But they write: “Just as electricity’s true potential was only unleashed when the broader benefits of distributed power generation were understood and exploited, AI will only reach its true potential when its benefits in providing prediction can be fully leveraged. For us, that points squarely at the role prediction plays in enhancing decision-making.” They go on to suggest that in many cases prediction will so change how decisions are made that the entire system of decision-making and its processes in organizations will have to adapt — and then AI will really take off.
Goldfarb was due to make a presentation about the book at an event at London’s Canada House last week, but was prevented from travelling for personal reasons. In his absence, a panel including an intellectual property and technology lawyer, an experienced chief medical officer and practitioners using AI debated the potential benefits and problems resulting from the development of the technology. Their insights added to the impression that we are still some way from seeing the true benefits of this powerful technology and — perhaps more importantly — still have the opportunity to ensure that we do not create the monster of popular fears. More joined-up legislation in such areas as IP and protection of individuals would help, but so would a greater awareness of the fact that humans are not totally powerless in deciding how things will turn out. As one participant pointed out, AI is not making decisions. It is helping managers make decisions.
But this does not mean that the ramifications are going to be limited to this. Only today, there have been press reports that a leading international educational qualification is to allow students to quote from work generated by ChatGPT. Just as they have had to become used to “digital natives” who have grown up with smartphones and their applications, organizations will soon be dealing with employees who see AI chatbots as as natural as calculators and spell-checkers. And they need to be prepared.