Why AI doesn't mean checkmate for humans

What can chess teach us about the marriage of man and machine? Author and Journalist Ian Leslie explores why blending human intelligence and artificial intelligence is critical to the future of our industry.

On May 11, 1997, Garry Kasparov raised his hands in frustration and walked away from the chess table, having been decisively beaten by IBM’s Deep Blue.

The world reeled in shock. An artificial intelligence had beaten a human genius. Did this spell the end of humanity as we know it?

As it turns out, it didn’t even spell the end of chess.

The following year, Kasparov founded a new tournament in which he invited teams to compete, each one composed of humans and computers. He called it Centaur Chess, named after the mythological half-human, half-horse.

Kasparov’s insight was that to view the relationship between humans and AI as adversarial is very limiting. What if each can make the other better?

In recent years, there have been big leaps forward in AI capabilities, and fears about computers taking over the world, or at least taking our jobs, have returned with a vengeance. The advertising industry, prone to anxiety at the best of times, is fretting about how to respond.

The answers fall, broadly speaking, into two camps. Those in the first camp are eager to hail our new overlords. They predict that as AI enables more sophisticated targeting and consumers outsource more decisions to algorithms, there will be less and less need for image-led branding. The sooner we recognise we are becoming a technology industry, the better.

The second camp says this is nonsense. The ad industry is about everything that computers can’t do: innovation, surprise, lateral thinking. Neural networks might be able to recognise existing patterns, but they will never invent truly new ones. Leave AI to the consultancies; we should be focused on keeping alive the ineffable flame of human creativity.

But perhaps there is a third way. We have a habit of conceiving of technological change like a chess match: a binary competition in which there is a winner and loser. We have inherited a centuries-old narrative about man versus machine.

History shows, however, there need not be any losers in this game. Time and again, technology that seemed to be about to make human labour obsolete has ended up making what we do even more valuable.

Perhaps AI is best seen, not a threat to our industry, but as a challenge; a raising of the bar. It will sweep away many of the tasks that currently keep us busy, particularly tasks that are based on existing patterns of behaviour, and which can easily be quantified and ranked. But that should in turn inspire us to be more creative, more empathetic, less predictable.

In 2005, a new chess tournament, inspired by Kasparov’s idea, pitted “Centaur” teams - AI + humans - against solo humans and solo computers. The grand prize was won by a Centaur team. The technology expert Nicky Case has highlighted the really significant revelation of the tournament: the winning team was not comprised of a human grandmaster with a super-powerful computer, but of relatively weak computers allied with amateur humans.

Why? The winning humans had worked out the best way to integrate their skills with the machines. That gave them the edge over teams that were, on paper, much stronger. Kasparov himself concluded that a “weak human plus machine plus better process will beat a “strong human plus machine plus an inferior process.

If agencies are to thrive in the era of smart machines, they need to do three things. First, put aside time to think strategically about the possibilities of AI rather than reacting to the hype. Second, identify the respective strengths of humans and computers as they relate to their unique business. Third, implement processes that lock the two sides into a symbiotic feedback loop of improvement.

The industry’s future belongs neither to technology, nor to human creativity, but to the marriage of both.

Ian Leslie is an author and journalist who writes for the Economist and the Financial Times on psychology, technology, and business. He is also a former agency planning director and now an independent brand consultant.

Ian will be co-hosting the How to be Human course on 22 November alongside Jon Leach. This course will explore how your agency can form a partnership with AI in order to move up the value chain. Or rather, how to be human – and win. 

Last updated 01 May 2024