Welcome To The Cambrian Era

Now, before you start thinking that I’ve entered a very dark place with this image, all I did, using Canva AI (other Gen AI image creators are available) was to enter the query, “AI Hallucination” and this was what came back. So there you are.

I, like many I imagine, know, or think I know, a reasonable amount about AI (who ever refers to it as artificial intelligence any longer?) LLMs, Generative AI and of course tools such as Copilot. But, on the basis that my Gen Z years are a long way behind me and I still use two fingers to type out a text message, in reality of course, I know very little.

I confess that it’s a slightly surreal experience, not to mention unusual, to attend a conference where I know not a soul. But this was technology as opposed to retail and at the ISG AI Impact Summit, I was there to learn.

And in this context, November 30th 2023, is a date which should be etched upon everyone’s memories, because that’s when OpenAI launched ChatGPT. Democratising Gen AI such that since then, 350,000 new websites and 15,000 startups in Europe alone - all with AI at their core, have launched. And now we have entered the era of personal GPTs, where you can create your own tool and your own capabilities.

“We're actually working where we think we can completely reimagine the engineering lifecycle, where it's not about waterfall or agile, but a brand new way to collaborate with AI in ways that we can't imagine”

Steve Hall, ISG

Over the last twelve months or so, I’ve heard a lot, an awful lot about AI from a retail perspective, at the many conferences I’ve attended, but this was the first time I’d listened to speakers who are immersed in the technology as opposed to an industry focus. 10.7 billion queries on OpenAI per week is still small compared with Google’s 22 billion searches per day; but all in all, it was quite startling.

As ISG’s Steve Hall described in his opening session, the world has seen a Cambrian explosion of generative AI applications in the last decade, with companies across industries rushing to get to grips with concepts like neural networks, diffusion models and natural language processing, and have encountered a wide array of risks and opportunities.

According to Hall, “It's teaching us to be more inquisitive, it's teaching us to be deeper, it's teaching us to explore things in a different way”. And one of the areas where this is having a huge impact is in coding and application development. “We're actually working where we think we can completely reimagine the engineering lifecycle, where it's not about waterfall or agile, but a brand new way to collaborate with AI in ways that we can't imagine”.

“The fast pace of the technology itself means that you can almost paralyse yourselves by indecision”

Mark Farbrace, Accenture

Now, all this sounds like something of a utopian dream, in the same way that when email and the internet entered our lives, we were told we would have so much more time for recreation (that aged well) but this time the change is on steroids.

But of course, as we know, if something seems too good to be true, it usually is. And whilst I wouldn’t wish to be a naysayer, just with every ‘new’ technology, a cautious approach, balancing the investment portfolio so to speak, is often the wisest course of action. “Rush in carefully”, in other words.

And this is the challenge facing every business right now; we’ve moved from the fun, personal experimentation with Gen AI into, “just how do we utilise this technology in a business scenario?”

According to Mark Farbrace of Accenture, “If it lasts more than three months as a delivery cycle, probably the tech world will have shifted by the time I get around to launch again. So I often think about Gen AI as a muscle and it's a new muscle that you're learning and kind of training”.

As I wrote back in January, AI, if we’re not careful can become the Sorcerer’s Apprentice, and Microsoft Copilot provided an example of that very recently when it accused a journalist of being an escapee from a psychiatric institution (and far worse which I won’t repeat here), the result of so called ‘AI hallucinations’, a scenario where it produces inaccurate or nonsensical responses to a prompt provided by the user.

This of course, is not limited to Copilot; however, it is a graphic example of why humans can never abdicate our responsibilities in favour of the algorithm.

However, none of this is likely to arrest the rapid growth of ChatGPT and its use across all industries. Google may still be the number one search engine by far and it might seem difficult to imagine a world without it, but there again, neither did we imagine we’d inhabit a world without Kodak.

Just as we migrated from analogue to digital, so we are in the midst of an even more seismic shift, because this time, the very essence of how we work and do business is fundamentally being re-written.