The human touch in an artificial world

It’s been a fairly stressful few weeks in our household, and last week my wife suffered a Ménière’s disease attack. It’s a horrible, incurable condition affecting the inner ear, which can leave her completely incapacitated for hours and needing several days to recover. Stress can make things worse, so we decided to spend a quiet weekend doing as little as possible and getting out into nature.

We headed out to Mamhead, a beautiful bit of woodland with wide, relatively flat paths, where you can wander around enjoying the great views and autumn colours without having to climb a mountain.

I also have a challenge for you that runs until the end of October: catch a falling leaf before it touches the ground (no cheating, it has to be naturally falling). We’ve been doing this for years, which means spending a disproportionate amount of time looking upwards and occasionally making sudden, inexplicable movements.

On this particular walk, I spotted one drifting down just beyond a man sitting peacefully on a bench. I immediately broke into a sprint, shouting “LEAF! LEAF! LEAF!” as I charged towards him, only to watch it land on the floor before I could reach it.

The look I got.

Down in the deepest shire of Devon, where people still occasionally say hello to complete strangers they pass on walks, I’ve been conducting a little experiment. I’ve noticed that you can often tell whether somebody wants to interact long before you’ve seen their facial expression, let alone exchanged a glance.

There’s something about their pace, posture and the general way they carry themselves that tells you whether a friendly hello will be welcomed with a greeting, a nod or a look suggesting you should mind your own business. My gut has been surprisingly accurate, although charging towards strangers shouting about leaves probably makes their decision considerably easier.

Have you been watching The Celebrity Traitors? Essentially a TV programme dedicated to proving how spectacularly wrong our instincts can be, with everyone confidently accusing entirely innocent contestants of being liars.

In the digital realm (not enough realms nowadays are there?), in the same intuitive way, we’re becoming increasingly familiar with the way AI writes, speaks and presents information. The sentence structures that shout “Claude”, the unmistakable feel of a ChatGPT-generated poster, or articles (like this one, perhaps) that manage to use an awful lot of words without saying anything particularly interesting.

We make a judgement about whether something is worth our attention just as quickly as that stranger on a walk from 100 yards away.

So you can see where I am going here, the question is, like the bumbling faithfuls on The Traitors, whether we’re getting any better at making the right judgement?

The bottom half of the internet

And yet, we’re all using AI, aren’t we? Or are we?

A friend recently described AI as being like teenage sex: everyone’s talking about it, not many are doing it, and those who are probably aren’t doing it very well.

Did you see the story from Mumsnet? A forum that has built its reputation on strangers sharing real experiences, asking questions (and occasionally having spectacular Jeremy Kyle-style arguments), found itself answering questions about its use of AI after an internal prompt appeared on the site.

Mumsnet backtracked and said that AI was being used to help suggest discussion titles rather than generate users’ contributions, but I’ll leave you to decide how easy it might be to have AI users and content on a forum that might benefit from the odd “emotional stir”.

Once people start wondering whether the person they’re talking to is real, trust becomes a rather complicated business.

And yet, while we’re becoming increasingly suspicious of who we’re interacting with online, Google is sending more of us towards places where those very conversations are happening.

For years, businesses have invested in carefully written, search-engine-optimised content to demonstrate their expertise and attract visitors to their websites. Now we’re increasingly seeing Reddit discussions appearing alongside, or above, those carefully crafted articles.

Imagine that, a forum where the answers might come from somebody sitting on their sofa in their pants, eating Wotsits and offering medical advice based on something that happened to their cousin in 2007.

What they will have, however, is a strong opinion.

Anyone who’s spent more than five minutes exploring the comments section of the internet will know that humanity doesn’t always present its best side when given a keyboard and a little anonymity, but I do think there’s something revealing about our appetite for these spaces, (and then the commercial reality that follows big groups of people).

And as a quick aside, for all the hurtful, racist, sexist and whatever-else-ist despicable behaviour, some of the most laugh out load funny things I’ve read have come from these gruesome depths, here’s a couple from memory:

I enjoyed whoever it was that announced: “If farting isn’t one of life’s great pleasures, then I’m not drunk enough.”

Another, following an injury to Declan Rice during the World Cup, suggested that he needed to be put in a bowl of mobile phones.

It’s the sort of beautifully ridiculous observation that catches you completely off guard, and which would almost certainly be removed from a corporate social media content calendar for not aligning with the brand’s strategic objectives.

And then there’s Jeff Goldblum…

Somewhere on Facebook there’s a page dedicated to posting exactly the same picture of Jeff Goldblum every, single, day.

That’s it. That’s the entire proposition. I was shown it a decade ago and when I thought of it a moment ago and found it, it’s still going.

No thought leadership, no carefully constructed customer journey, just Jeff Goldblum, turning up every day, looking exactly as Jeff Goldblum looked yesterday…And people love it.

There’s something wonderful about a community built around something so gloriously pointless, particularly when you consider how much time businesses spend trying to create engaging content. The repetition becomes the joke, and everybody following it is in on it.

The internet might have been built to connect computers, but humans quickly turned it into somewhere to connect with each other.

And so the chasm widens and deepens. On one side, we have these wonderfully human spaces filled with niche expertise, shared experiences, unusual interests and people being spectacularly themselves. On the other, an ever-growing mountain of generic, repeatable AI content.

Something doesn’t feel quite right

And unsurprisingly, a recent survey suggested that 62% of UK consumers trust marketing content less since brands started using AI, while 57% believe they can usually recognise when something has been generated.

That second figure is funny because believing you can recognise something and actually being able to recognise it are two very different things, and the number is very low for something we may be somewhat over-confident with.

We might be dismissing perfectly genuine human communication while happily consuming AI-generated material without giving it a second thought.

AI has made it possible to produce more content, more quickly and at a fraction of the cost. And I think that’s because we’ve confused the ability to produce content with having something worth communicating.

What makes the comments section interesting isn’t the quality of the writing. It’s the personality, the unpredictability and the context in which somebody has said something.

What makes Reddit useful isn’t necessarily the authority of the person answering. It’s the possibility that somebody has actually experienced the problem you’re trying to solve.

And what makes Jeff Goldblum entertaining isn’t the picture (well, it is a bit) but mostly it’s the ludicrousness of the daily promise and why somebody keeps showing up and posting it.

These are all things that depend on context, and context is something humans are extraordinarily good at intuitively understanding.

Think back to that person walking towards you. You don’t need them to explain their mood, their intentions or their relationship with the world around them. You’re drawing conclusions from a collection of signals that mean something because of everything you’ve experienced before.

AI, on the other hand, doesn’t have that lived experience, and as I mentioned last month, it doesn’t know what it’s like to have a difficult conversation with a customer, make a mistake on a project, build a business through a recession or spend an afternoon wondering whether you’ve made the right decision.

Unless, of course, we throw some of that context into the mix.

Finding the human in the machine

Follow this thinking through and there’s a growing temptation to divide everything into two categories: the genuine stuff created by humans and the artificial stuff produced by machines. This, to me, feels just as black and white as all of history’s greatest mistakes.

I don’t particularly care whether somebody used AI to help organise their thoughts, improve their writing or research a subject. What matters to me is whether they’ve contributed something interesting, whether the information is reliable and whether there’s a genuine reason for it to exist.

Ask a writing tool to produce an article about sustainable business and you’ll probably get something perfectly reasonable about reducing emissions, engaging stakeholders and embracing innovation. It’ll be grammatically correct, professionally structured and could have been written for or by almost any business in the country.

But what if that same tool understood the organisation properly?

What if it knew about the projects they’d delivered, the mistakes they’d made, the customers they’d helped and the unusual ways they’d solved problems? What if it understood their values, their opinions, the questions they get asked every week and the things they believe their industry consistently gets wrong?

Suddenly, there’s something rather more interesting to work with.

And this is one of the ideas we’re exploring through the development of Whyso and the Vu dashboard.

We spend a lot of time getting under the skin of businesses, understanding their ambitions, their customers and the things that make them different. We combine that discovery process with information about where they’ve come from and how they’ve performed, but so much of the valuable knowledge we uncover ends up scattered across meeting notes, documents and people’s heads.

What started as a way of bringing those things together is evolving into something we’re calling a context engine. Not just a record of how a business has performed, but a richer understanding of why it does what it does, what it’s learnt and what it has to offer.

Because if AI is going to help tell your story, it needs to know what your story actually is.

And perhaps there’s a bigger question here about what we’re actually trying to achieve with all this technology.

We’re surrounded by conversations about how AI will transform business, make us more productive and help us grow, but I wonder whether we’re spending enough time asking what we want that growth to achieve.

Surely the point of making businesses more successful, creating better technology and finding more efficient ways of working is to improve people’s lives? Starting with the people closest to us, extending into our communities and, hopefully, making a positive difference a little further afield.

Otherwise, what exactly are we optimising for?

There’s still an exciting opportunity for smaller businesses to make better use of the knowledge and experience they already have, and whether that’s through AI, better systems or simply taking the time to understand their customers, the technology should be helping us achieve something worthwhile.

For all the exciting things we’re exploring with AI, the businesses we’re building and the problems we’re trying to solve, none of it means very much if we’re not making life better for the people around us.

Last weekend, that meant forgetting about work, getting Jen out into the woods for some fresh air and spending a few hours enjoying the autumn colours.

And occasionally making a complete idiot of myself in front of strangers.

And I still haven’t caught that bloody leaf…