The late Stephen Hawking once said that success in creating Artificial Intelligence “could be the biggest event in human history”. Now that we are daily witnessing generative AI technologies experiencing huge leaps in capacity and delivering ever-increasing numbers of applications, it seems obvious that he was right. CEOs need to engage actively with what is unfolding in what is clearly a field with truly revolutionary potential for many aspects of business and for our lives more generally.

For many of the CEOs I am working with that is already obvious. The key question is how should they respond? How and where should they lean in, and where should they delegate? This uncertainty is not at all surprising. After all, for as long as we can remember CEOs have been encouraged to focus more on “being, rather than doing”, to empower their reports and allow interesting ideas to bubble up through their executive team. It is fair for them to ask whether AI is in the end any different from other technologies in this respect. Is it just posing a new set of strategic choices to be navigated in the usual way, with their team doing much of the detailed work? Or is there something different about what is happening in this field that requires a bespoke approach? My strong sense is that this needs to be considered by CEOs as both a practical strategic challenge, and also as a profoundly emotional one. If this intuition is right, then quickly building a good sense for the fundamentals of what generative AI does, and how it does it, will be particularly important for CEOs if they are to remain on top of those challenges, both personally and as a leader.

Agrawal, Gans and Goldfarb make clear in their helpful overview, Prediction Machines (Harvard Business Review Publishing), that there is good reason for CEOs to get more directly involved in thinking through how AI will shape their future than is the case for some other challenges they might be facing. The most powerful argument the authors make is a disarmingly simple one. Artificial Intelligence has the potential to upend existing businesses even more profoundly than other recent developments like the internet. They give the hypothetical example of a business like Amazon. They imagine a world in which AI has got to the point where it can help Amazon anticipate customer’s needs so accurately that they decide to deliver goods proactively to their customers before they have even ordered them, and guarantee free returns if they get it wrong. The authors point out that this would be a revolution in how Amazon approaches their market. And, just as powerfully, if a competitor got there first and succeeded in making this approach work, it could be a major threat to Amazon’s business if they were not able to replicate it.

In this sense, AI is an inherently fundamental strategic development that demands significant attention from the CEO. Agrawal, Gans and Goldfarb indicate that Artificial Intelligence is likely to pose some critical choices like this in almost every business of scale sooner or later. But they also point out that there is another key feature that is specific to AI that heightens the threat to existing incumbents in any given market. The machines involved usually need significant time interacting with customers or other key elements of the business to learn how best to deliver what is needed in any given business application. This has important consequences in terms of timing. New entrants who have recognised a strategic reinvention opportunity, and have started gathering the relevant data from customers earlier than established players, may well have the potential to steal the march on those incumbents. Given the time needed for AI to learn how best to make sense of what is often quite complex data, fast follower strategies may simply not be fast enough to catch up in this context, no matter how much money is ploughed into them.

The implication seems to be that the smart CEO should initiate early and quite boldly-scoped discussions on AI and its potential impact on their business, and be closely involved in those investigations from the start. The scope and speed of change may well mean that external consultants are a natural part of determining how best to respond. So often they are hugely helpful in speeding up the work of an already busy executive team. However, in this case they would need to be encouraged to structure a novel process that allows for an exceptional level of CEO/Executive team ownership and participation. Promoting “self-discovered logic” through close teamwork with clients has, of course, been a mantra of such consulting for half a century at least. But it would be hard to deny that for most hard-pressed CEOs their involvement is usually confined to scoping the work, understanding the analysis and determining decisions. In the case of AI, it seems that the depth of the changes involved, and their implications for CEOs’ people, mean that there will be real value in leaders familiarising themselves more actively with the fundamentals of the technology and how it works, and not just understanding what experts say are the implications.

If this is right, they will need to shift some of the boundaries of where and how they get involved from the outset. It will be essential to create teams that are both highly qualified to explore future possibilities and that are really exceptional at collaboration, with deep levels of mutual trust, to ensure full ownership and understanding of the outcomes by the CEO and their team. Once the key decisions have been flagged and the risks involved weighed up with the Board as needed, the CEO who has been actively involved in the work can then make timely and well-calibrated decisions on what may often be high stakes changes to the strategy going forward. The logic suggests that it is then - and only then – that the CEO can start delegating much of the implementation in a more conventional way.

So far so good. Yet many CEOs may still feel that in the end this is a set of important strategic choices that are in essence not very different from many others that they have seen. They may still be inclined to treat the emergence of AI with respect but in the end as something whose implications can be managed within their current ways of working. There is to my mind another side to this question of CEO involvement. This goes beyond staying close in order to respond in the most creative and agile way to what is a potentially existential strategic threat. It relates to the leader’s role in providing an environment of psychological safety within which their people can perform. If Hawking was even half correct in his sense of the potential of AI to transform human existence, then we are at the foothills of a mountain of change that we cannot see clearly yet, but which – given the pace of current developments - seems likely to spring suddenly into view. We cannot yet know what the societal, corporate or individual impacts of the thought of scaling that mountain will be. They will of course vary immensely in our highly complex and varied world. One of the more perplexing aspects of this revolution is how little some of those initiating it seem to be thinking through its longer-term implications. But we can expect that many traditionally valued roles may look likely to be rendered obsolete in pretty short order.

It is this combination of a sense for the power of the changes hitting the world, with the lack of clarity about what they will mean for us, that has the manifest potential to trigger widespread fear, possibly on a scale and to a degree that we have not witnessed in society for a long time. The implications for the CEO in this scenario are much more than defining the right strategy. They will need to support their team (and potentially other important stakeholders) by being a source of trusted advice, ensuring they remain grounded by giving them a clear sense of direction based on knowledge and insight, rather than bravado or false certainty. The suggestion here is that the CEO’s level of familiarity with the fundamentals of artificial intelligence may dictate how much authority they can bring to the effort to help ground their people as they are faced by so much change. The more they are themselves grounded in a deep engagement with the options ahead, and how to judge which ones are most likely to be important, the more they will be able to transmit a sense of purposefulness and clarity to their immediate team and the wider organisation. This may simply be to reassure them that the best evidence thus far is that their current operating model will continue to be effective. Or it may be to announce a complete reshaping of the company, but with a really well-considered approach to managing the implications for any people in the business who are impacted. Either way, the more deeply this new direction is felt and embodied by the leader of the organisation in question, the more powerfully they will be able to shape and channel their people’s fear.