Edition 5 : The Serious Issue of the Even Bigger Bang

AI: The even Bigger Bang

Starting in the mid ‘80’s I worked in The City at the heart of the chaotic whirlwind of ‘The Big Bang’.

You could literally see a new world evolve in front of your eyes; elder gents in pin stripe suits got ever rarer whilst big shiny new buildings became ubiquitous. All of it driven by new tech as data effortlessly leapt from hard copy to floppy to real time.

The results were staggering; jobs in finance went from 90,000 to 325,000, the number of foreign banks went from under 40 to over 160 and, in just three years LSE share turnover went from £700 million to just over £2 billion.

For my part I went from paper based data analyst to training very tech-resistant old school brokers who for over a century had practised the frenetic “Sell, Sell, Sell! Buy, Buy, Buy!” of Open Outcry and now had to learn to utilise the far faster, far more accurate and far more efficient but much less theatrical electronic dealing systems.

Many of those Dealing Rooms were still run like an Edwardian public school common room. I had one chap casually set fire to the copy of the FT his colleague was engrossed in, another never stopped playing poker with his rival at the opposite end of the dealing floor, using hand gestures and a runner to distribute cards, chips and winnings.

They were not alone in their resistance, there were major institutions that remained reluctant to change even in the face of inevitability. I spotted a cushion on the sofa of the MD of a now long defunct private bank, it read ‘old age and deceit will always overcome youth and enthusiasm’. Wise words or gallows humour?

But even though it was in a permanent state of upheaval The City as a whole knew it had to do what it had to do. It had over decades become disastrously complacent and was fast being wholly eclipsed by both New York and Tokyo whilst Frankfurt, Paris & Amsterdam were poised and eager to snatch its crown.

The Big Bang was planned and executed in remarkably quick time even though the outcomes were uncertain, there was a take your medicine method amidst all the madness.

But today with the advent of AI no such planning has been instigated which is a worry as this is another much much Bigger Bang and this time it’s on a truly global scale.

What can we do?

Even if we halted all work on AI today estimates say is it’d be at least two years before we worked out the scope of capabilities of what Chat GPT4 is capable of.

So we’re aware this is not some fad or flash in the pan and who knows, you may ignore it and hope that in the end AI won’t really change things that much. If you do I can guarantee that people around you won’t and if they work for you, they’ll find covert ways to use it without you knowing about it.

You can try and outlaw it from your business, which is really just an extreme way of ignoring it only this time you want everyone else to ignore it as well. Good luck with that as we both know your people will still continue to use it (on their phones, at home, at lunch, in the loos etc) and you still won’t know what they’re doing.

You could trust traditional business wisdom which would be to implement some kind of internal centralised system. Which means you get to control it and you and the Senior Management Team can make the decisions about how AI will be used- all be it based on your limited experience and understanding.

This way seems understandable as it’s the usual business standard for what is done in the face of new tech; you shop around and eventually you buy in the expertise along with some kind of pre-packaged system which allows you set policies on use, monitor usage and keep your data ring fenced.

Which should work because it’s worked with every previous tech advance, it must work otherwise why would these systems cost a fortune to install, integrate and then train everyone to use?

Well, remember the saying about how Generals are always fighting the last war?  From a business angle buying in a systemised centralised AI is like digging a trench in response to the other side having just unveiled its new fleet of Challenger 2s.  

Because AI simply isn’t in any way built for doing this - at all. If AI is anything it’s the complete opposite.

There’s no advantage. Chat GPT is the most advanced AI available, and it is free for everyone via Bing, or for a small monthly fee from Open AI. The size and nature of a business offers no special access whatsoever to anything better. None at all.

But let’s say you go ahead. Now what? What are you hoping to achieve?  So far there’s no evidence or reason to believe any organisation anywhere (including the people who built the AI in the first place) has or are going to ever be experts at understanding how AI can best help anyone they employ with their problems or task. They’re more likely to be really bad at knowing let alone figuring out the best uses for AI.

This is because AI is designed to be extremely good at being ‘present problem context centric’ which means it’s the individual or small team that are always going to be best placed to know the problems they need to solve. It is those at the actual coal face who will always be best placed to find uses for what AI can do.

Meanwhile your brand new system is stymying everything in every way. You have to monitor its use and your people will all know they are being monitored, and they’ll face punishment if they use it in some as yet to be defined wrong way. So, what will they do? They are smart and they’ll quickly find workarounds mostly by using the freely available AI (at lunch, at home, in the loos) and if need be then transferring the results to your system which means you have just gone to great lengths to install what is ultimately a very expensive filing cabinet filled with things that are not actually very useful, interesting or in any way reflect the kind of hugely powerful things AI can achieve.

It's a bit like the WW2 scenario where the Special Operations Executive in Washington allegedly had a state-of-the-art underground bunker with rows and rows of sleek, steel filing cabinets – all of them empty. Meanwhile Bletchley Park was beavering away in a bunch of draughty Nissen huts stuffed to the gunnels with shoeboxes overflowing with real useful usable data.  

In this context I have to say AI is a bit of a Great British dream scenario: AI is an incredibly useful but complex and challenging problem that is absolutely crying out for practical ‘lash-ups’, ‘workarounds’ and ‘whatever works now’ rather than idealised theoreticals. There may be good reasons our own Alan Turing was the (extremely bright) spark that kicked AI off.

So, to opt for any kind of One-Size-Fits-Nobody top-down approach to AI is fighting the last war purely because that’s the war you know how to fight.

AI is Not Jam

The problem is AI isn’t jam. By trying to label AI as being ‘akin to’ or ‘like’ some or other technology you know about completely fails to see just how startling different it is.

We can’t yet put precise labels on AI because we don’t yet have a label that adequately encompass let alone describe what it can do.

To put it another way AI has so many potential viable labels that no single one is even barely adequate at describing what it does with any accuracy to makes the label valid.  

The Buddhists seem to have captured it quite well; a philosophy as deep as it is ancient boils down to the premise: 

“Don’t apply labels to things because labels sets a thing in stone, and you come a cropper when that label turns out not to be an accurate reflection of the thing you’ve labelled.”

Now obviously, I’m very much simplifying 2,000 years of wisdom for the sake of convenience, but you get the gist.

Being un-label-able is AI’s great strength - a genuine feature and not at all a bug - as it allows AI to act as a force multiplier for any individual or small team enabling them to do a tremendous amount of work, but it is a different sort of work.

AI is like having an incredibly keen, bright and ever eager unpaid Intern who absolutely laps up doing all (and I mean ALL) the tedious repetitive tasks we probably loathe having to do -  by doing this it frees you to concentrate almost entirely on the interesting, innovative tasks you went to work for in the first place - and as an extra enormous bonus it allow you to hugely multiply the effect of those tasks and the results gained.

But In doing this the very nature of what we mean by work faces a genuine paradigm shift in ways that can seem uncomfortable, unknowable and risky. What we do know is that our current systems really aren’t built for it at all and that this new Ubiquitous Big Bang we are in the middle of needs to be urgently not just looked at but urgently acted upon.

Right now, the most advanced uses and discoveries in AI’s potential is being done by individuals like me. That’s not a brag, humble or otherwise it’s just people like me are (like AI) already ‘decentralised’ & unconstrained by specific business models and right now there’s simply no way for companies to harness the power and creativity over AI without finding and adopting a broadly similar ethos.

Only innovation driven by individuals or small teams can transform work in the way it has to transform, because only individual workers or small teams can experiment enough on their own tasks to learn how to use AI in transformative ways. Sadly, empowering workers is often not feasible with our current old school top-down solutions.

But I wouldn’t be worth my salt if I didn’t offer a way forward and in my next blog post I’ll be doing just that.

In the meantime if any of the above has struck a chord I offer a free thirty minute discovery call where we can discuss anything you want around AI and your business just click below to book.