Ubiquitous yet hated – what does the triumph of PowerPoint teach us about Generative AI?

19th October, 2023

The aesthetic of our age was shaped in Paris in 1992, in the Hotel Regina. The occasion was carefully stage-managed by a team of technicians fussing over a huge colour projector that cost as much as a small house. The big unveiling came when Robert Gaskins, a Microsoft software engineer, walked up to the lectern, plugged his chunky laptop into a video cable and began showing PowerPoint slides in full colour, straight off his machine. The applause was, according to Gaskins, “deafening”.

There were visual aids before 1992, of course. At the high end, there were computer-co-ordinated slideshows in which dozens of projectors were choreographed to fit with music, script and each other, producing spectacular results at extraordinary expense. 

The mid-market was a monochrome or colour transparency placed on an overhead projector (OHP). In the heyday of the OHP, more than 2,000 were sold in the US every week. (For a detailed and delightful history of visual aids, I recommend Ian Parker’s “Absolute PowerPoint” in The New Yorker in 2001 and, more recently, Claire Evans’s “Next Slide Please” in MIT Technology Review.) 

Or there is the literally old-school approach: write on a blackboard, whiteboard or flip-chart.

Gone, all gone. These rival visual aids have been driven to near extinction by PowerPoint and Keynote, made by Apple.  This is odd, since few people love PowerPoint. Hotel Regina is a five-minute walk from the Louvre, but PowerPoint is a universe away from fine art. Gaskins and his colleague Dennis Austin, who passed away earlier this month, managed to create a product that was cheap, ubiquitous to the point of inescapability and widely reviled.

How did bad PowerPoint triumph? And what can we learn from that victory? One lesson is that when it comes to technology, we’re lazy. We reach for the nearest familiar tool without thinking about whether it’s the right one for the job, or even thinking clearly about what the job is. Are we trying to think through a problem? Get a discussion going? Show people that worth-a-thousand-words picture? We skip that vital contemplative step and load up a slide template instead.

Because everyone can use PowerPoint, everyone does. That is how highly paid managers, engineers and lawyers end up fussing about fonts and colour palettes.

PowerPoint is not to blame for this, any more than I should blame a Swiss Army Knife for poor results if I rely on it when putting up some shelves, rather than using a full set of tools. The fault is our tendency to grab whatever is within reach. 

One can see this by observing much the same tendency in our lazy, indiscriminate use of PowerPoint’s sibling, Excel. Type “SEPT1” or “MARCH1” into Excel and the software will automatically convert those inputs into dates. That is usually fine, but unfortunate if you were a genetics researcher referring not to dates, but to the genes with those names. The gene autocorrect problem was spotted nearly 20 years ago and appears to be getting worse. The proportion of genetics papers with autocorrect errors was estimated in 2020 to have reached 30 per cent. The Human Gene Name Consortium decided to rename the genes in question, wisely accepting that this would be easier than weaning researchers away from Excel. 

Compared to the way that generative AI will be similarly misused, such problems may come to seem small. We’ll ask Google’s Bard AI to sketch out an argument or Dall-E to draw us a picture, even if the results are often patchy. Why? Because at that difficult moment, when we’re staring at a blank page and wondering what to do, these tools offer escape. PowerPoint once included an “Autocontent” feature. That displays considerable insight: we humans will seize any technology that might liberate us from the tiresome need to think for ourselves.

In Thinking, Fast and Slow, Daniel Kahneman observes that when faced with a difficult question, we often subconsciously find an easier question that seems relevant, and answer that instead. This can be a useful approach, but the danger is that this process of substitution is so effortless that we may not even realise we have done it.

In the world of presentations, PowerPoint often plays a role in this subconscious switch. We are faced with a hard question: when standing up in front of an audience, what do I really want to communicate and how should I do that? It is vastly easier to ask, what are the first 50 bullet points that come to mind when I think about giving a talk? And then to pretend to ourselves that the two questions amount to the same thing.

The results are tedious, overstuffed talks in which the speaker’s notes are plastered on the wall behind them in advance. Better to print those bullet points on to 3x5in note cards, but that would defeat the subconscious goal of allowing the speaker to step as far away as possible from the centre of attention. Many presenters wish they could simply vanish. Using PowerPoint like this, they might as well. 

I don’t love PowerPoint, but as a technology there is nothing much wrong with it. It can do pretty much anything that you can do with a computer-choreographed barrage of slide projectors, and much more besides. And it can do it more flexibly, more reliably and much, much more cheaply. 

Yet that is the trap. A great talk starts with a message. Everything else — whether a joke, a story, a statistic or a picture — should be chosen to support the message. It’s always been easy to forget that. In a world of PowerPoint on tap, it can be impossible to remember it.

Written for and first published in the Financial Times on 22 September 2023.

My first children’s book, The Truth Detective is now available (not US or Canada yet – sorry).

I’ve set up a storefront on Bookshop in the United States and the United Kingdom. Links to Bookshop and Amazon may generate referral fees.

Pin It on Pinterest

Share This