Thinking beyond AI

Complex systems may repeat themselves, but not predictably, so patterns can prove to be dodgy guides. Technology easily makes us forget what we know about life – but it is in the life of an organization that leaders must lead. Social context, embodied experience, wisdom over time: None of these are encoded in Large Language Models (LLMs). Uncertainty is one defining quality of human life, poorly represented in AI. All models make assumptions about what good looks like; in technology, this is usually assumed to be efficiency or utility maximization. But these qualities carry risks when too tight margins leave companies without capacity to ­respond to shocks and surprises. In ­addition, ­endemic uncertainty requires that not all choices be defined by the same values, but in LLMs what those values are is frequently opaque.

AI also has an inadequate sense of context, which keeps shifting. There is no data from the future, so AI must work with data from the past. That doesn’t mean it’s useless, but it does mean that it cannot reflect unpredictable shifts in attitudes, possibilities and values. It’s better seen as a provider of ­options than answers. So it is incumbent upon leaders to understand how, when and where AI is useful and how to interpret and question its outputs. As tempting as it might be to grasp at the ­machine-generated “answer,” to do so isn’t just an abnegation of responsibility, but a failure to understand leadership itself.

The gold standard of decision-making is the capacity to make choices that even those who don’t agree with them can understand. But how do you explain decisions you haven’t made – and don’t ­understand yourself? Active thinking, connecting knowledge to the world, is fundamental to trust, timeliness and legitimacy. Being able to think without banisters, to have a conversation with oneself, free of rigid ideologies, assumptions, habits and dogma, is critical for the life of any organization. It is what stakeholders look to leaders for. This may be one reason why, for the last two years, the World Economic Forum’s Future of Jobs Report insisted on the ­importance of both critical and creative ­thinking. More than answers, the art of asking great questions ­becomes paramount to analyze the meaning and consequences of AI “answers.”

Effective questions don’t presuppose binary ­answers; like a great surgeon, a great question will open up a body of contingencies and possibilities, probing for options and risk. Critical thinkers work from the assumption that there is always more to find, always something missing. This can’t be done alone; leadership thinking still requires the skill to think together, assembling the appropriate array of knowledge, experience, expertise and perspectives around questions that matter. That also requires the capacity to listen and to interrogate. AI, which is morally opaque, demands ethical examination: Without excavation, decisions can be made but their legitimacy will be fragile.

But why the emphasis on creative thinking? Technology frequently renders thinking, products and services generic, undistinctive and uncompetitive, making innovation vital. With AI, strategy becomes, more than ever, an act of imagination. Creative people are often ahead of their times; hence avant-garde. They start from curiosity, noticing anomalies and quirks, with no agenda, cultivating a well-stocked mind. Over time, random data points coalesce around something – an idea, image or sound – and creatives often start work with no plan.

The self-efficacy to find sense and meaning along the way, to invent through experimentation and through discovery, is a rich, adaptive form of sense-­making that develops initiative, originality and ­resilience. Such aptitude, together with a marked reluctance to repeat themselves, enables creative people to respond productively to the ­unexpected and to drive innovation. Applied to problem-solving, such capabilities may be idiosyncratic, but are the best (and perhaps the only) ­antidote to the generic ­nature of ­technology-driven output. It’s worth ­remembering that AI was once touted as a quick and easy answer to drug discovery. But ­the sheer complexity of human biology has disappointed investors.

Leadership once was increasingly ­technocratic. Now AI challenges us to contain paradoxes: to match urgency with patient ­questioning, ­focus with mind-wandering and skepticism with faith, the ­optimism all leaders must have that they, and those they lead, bring unique human skills to urgent problems and inventions. This shift won’t happen by magic. It will take some thinking …