CEOs must soldier on even as AI anxieties loom
Top executives need to use AI to empower employees rather than replace or monitor them.
Chief executive officers keep getting new jobs piled onto their shoulders. The economic rise of China since the 1980s meant that they had to become Sinologists. The twin populist shocks of Brexit and Donald Trump's presidential ascension meant that they had to think about the capitalist system's legitimacy. Now, with the next tech revolution at warp speed, they are having to become experts on artificial intelligence. It's almost as if they're underpaid.
During the China era, CEOs tried (at the very least) to master a few words of Mandarin. Today, they're sprinkling AI-speak into their conversations: foundation models, large language models, hallucinations and the lot. And they're desperately boning up on the latest books and obediently trooping to classes run by consultancies such as McKinsey & Co. and Boston Consulting Group. Gonzalo Gortazar, CEO of CaixaBank SA, sums up the mood in the C-suite: “Generative AI models surprise, impress and scare us, all at the same time.”
The speed of change is neck-wrenching. ChatGPT was released less than a year ago, on Nov. 30, 2022, reaching 100 million users in just two months. An International Business Machines Corp. survey of 3,000 CEOs found that 43% said their firms were using generative AI to make strategic business decisions, and 75% believed that it would eventually give their companies a competitive edge. A McKinsey survey found that a third of respondents said their companies were regularly using AI for at least one business function.
The technology has been developing quietly for years. But it was the arrival of generative AI — capable of producing text, images or other media — that seized public attention and the C-suite by the throat. Generative AI possesses general ability rather than just highly specialized ones. It is creative rather than just passive. It has learned how to speak to people through ordinary language. The cherry on the cake: GPT-4, released just four months after ChatGPT, was significantly cleverer than its predecessor.
The pressure on CEOs from various interest groups is massive. The IBM report shows that 66% of board members want to accelerate the adoption of AI compared with 10% who want it to slow. Questions about the impact are dominating corporate earnings calls.
You might expect some employees to be more nervous. But a survey by Fishbowl, a professional social network, of 11,700 workers, including employees at the biggest tech companies, suggests that they are adopting AI even faster than their bosses: 43% of respondents said that they used AI to do their work — and 68% of the 43% said that they hadn't told their bosses that they were doing so.
But the cry of “faster please!” will only get louder.
Meanwhile, consultants and some tech companies seem to be doing their best to stir up fear and greed. McKinsey claims that generative AI can add $2.6 trillion to $4.4 trillion “in value” annually across the 63 different industries it analyzed. (By comparison, the UK's entire GDP in 2021 was $3.13 trillion.) Alphabet Inc. CEO Sundar Pichai in 2018 said, “AI is probably the most important thing humanity has ever worked on. I think of it as something more profound than electricity or fire.” The message in a nutshell: Join the revolution or become roadkill!
The very thing that makes generative AI so impressive — its ability to scrape the internet for vast amounts of data and then answer questions in human-friendly tones — also makes it dangerous, raising ethical and practical questions. CEOs have little choice but to turn themselves into data experts — or at least data statesmen. Then there's the old problem of bias in a new form: What happens if the AI discriminates against people based on race or gender? Executives also must deal with what has been christened “hallucination” — AI's ability to present false narratives as if they are incontrovertible truths.
AI also supercharges the ever-thorny problems of privacy and security. The machines can ingest others' proprietary material (see: Getty Images) as if we live in a libertarian paradise rather than a world in which information is governed by strict laws of copyright. They can invade people's privacy by publicizing data that can be traced back to individuals or compromise corporate security by sweeping up spyware. Deepfakes make it easier for hostile interests to hijack a CEO's identity to deliver market-shifting and reputation-destroying announcements.
Management consultancies are all too keen to offer advice on how top executives should reorganize their firms to minimize such disasters: They need to revamp governance structures to keep a Sauron-like eye on generative AI — possibly by creating a special task force of both specialists and generalists. They should establish new safety measures in case of a “malfunction.” They need to improve their own familiarity, perhaps by incorporating more tools into their daily routines and increasing AI-savvy people in their workforce and on their boards.
The consultants are equally fulsome about how to maximize the upside of AI. Focus on your company's traditional competence first instead of pretending that you're a tech firm that just happens to make galoshes. And ensure that you are making the greatest use of your proprietary information rather than relying on information that is in public circulation. Procter & Gamble Co. has been collecting data on consumer goods and how to sell since 1837. Home Depot Inc. has a vast collection of material on real-life home improvement projects.
All very sensible. But the best CEOs will go further and ask even bigger strategic questions. The current gold rush raises the possibility that companies will use AI to replace workers rather than empower them, thereby destroying morale, frustrating customers and embedding second-rate performance. In their new book Power and Progress: Our Thousand-Year Struggle Over Technology and Prosperity, Massachusetts Institute of Technology economists Daron Acemoglu and Simon Johnson point out that corporations systematically underestimate the ability of humans and overestimate that of intelligent machines. Humans may not be as fast when it comes to reviewing vast quantities of data, but they are unique in their capacity to bring intuition and empathy to their jobs. The result of overreliance on AI could well be more “so-so innovation” of the kind that already leaves us fuming at automated voice responses or standing in a state of existential despair in front of retail checkout machines.
A Goldman Sachs Group Inc. report concludes that generative AI tools could affect 300 million full-time jobs worldwide. Coca-Cola Co. CEO James Quincey boasted to a conference at Yale University that, based on the results of a marketing campaign featuring AI-generated ads that proved indistinguishable from human-produced ones, the days of companies like his spending billions on ad professionals could be over. While most innovation in the past has impacted manual or low-skilled workers, knowledge workers — particularly those in law, administration and the media — are in AI's crosshairs. Hollywood writers are already on strike over, among other things, worries about seeing their jobs and residuals disappear.
It may sound self-serving to raise the question of worker displacement when so many of us celebrated the advance of automation elsewhere. Why should educated Luddites be spared the productivity-boosting march of progress? There are prudential reasons. Western societies may not be able to tolerate the hollowing out of the professional middle classes following the gutting of the lower-middle and working classes. A glance at the history of both Bolshevism and Nazism shows that knowledge workers have a particular talent for leading vicious revolutions.
But there's also a practical incentive: The CEO's most important job is to steer their companies in a human-centric direction — to turn AI tools into servants rather than masters and free workers to employ their most human qualities: originality, empathy and, above all, judgment.
Adrian Wooldridge is the global business columnist for Bloomberg Opinion. A former writer at the Economist, he is author, most recently, of “The Aristocracy of Talent: How Meritocracy Made the Modern World.”