Generative AI will scale mediocrity unless you look after the soil

‍What can we learn from the past about our future?‍

The scale of the industrial revolution was unparalleled for the time, the impact on agriculture is still unfathomable to us today yielding a one thousand times impact on our output it sent a shockwave across society. Things were changing and they were changing fast. Our society from x to y scrambled to adapt to what would be a new world.

Suddenly machines were amongst us, ploughing our fields and moving our goods, the land that once offered so little was turned into an efficient, high-performance beast. A litre of gasoline became more valuable than the strongest of workers, completing what used to be 30 days of work in minutes and with far less complaints. Humanity had achieved something truly great, food production that could supply the globe… at last we could rest.

You don’t need to be an economics major to know that all this increase in supply (and stable levels of demand) would begin to drop prices as people were able to live more comfortable and satiated lives than ever before. Of course, the story does not end here… a monumental shift was underway and not everyone enjoyed the fruits of this change to the same degree. Gaps began to emerge, especially the one between the rich and poor. And this was not the end of it, our burning desire for more scale, more efficiency brought about monocultures. Destroying diverse habitats and creating huge areas of land pumping them full of fertiliser to drive the growth of one plant… the issue with this was you now had an entire ‘ecosystem; that could collapse to a single infestation. This new weakness, birthed from the wonders of the industrial revolution drove the need for pesticides. AND to top it all off the food no longer gave us what it used to with nutritional value deteriorating massively. The strength and richness of the soil were reduced by all our interventions. In essence, and it should be no surprise, our mentality of hyper growth at all costs left us with some thorny problems.

Now… what does all this negativity have to do with AI? Well we are simply suggesting that there may be some lessons for us to consider when determining how we create the best possible A.I. future, as one thing is sure, it will be there. In particular we have focused on two themes:

1. What can we learn from the negative impacts of scaling automation and how do we mitigate this to make this revolution a little less problematic?
2. How do we mitigate the societal and organizational growth pains?

‍

Making it make sense

Now just like the industrial revolution, even more so, we need to have a conversation on how to oversee this subject. However it doesn't make a huge amount of sense to talk about Gen AI in its general sense… its impact is too vast.

At Leapfrog AI, we distinguish three types of Gen AI applications. First, Gen AI interprets and exchanges large amounts of data between tools and people. Second, it does creative production work, where 90% of design can be automated, and the work is personalised and localised at scale. Third, Gen AI does creative, conceptual, and strategic thinking. The latter can mostly be impacted by the negative effects we describe in this article.

‍

A desire for scale with a rich personalised AI machine

It’s all well and good that we can 1000x our output in creative production work… suddenly 10000 digital ads, localised to 22 countries is possible at the click of a button (a person should never be doing this anyway)

It does become a bit more problematic when use Gen AI for the work that requires a little more conscious thought. Imagine asking Gen AI (and it's not hard too) to apply marketing models for a client strategy proposition. Sure, it's able to create a plan but soon enough every single strategy deck looks the same and we all begin to morph into one lazy AI driven machine. So sure, we can produce more faster, but more of the same doesn’t drive us to quality and a higher level of being… which isn’t too much to ask for is it?

The more you can direct your own AI, the more your organisation will thrive. The more you can train AI on your style, your uniqueness, your tone of voice (TOV), your best briefs, your magical insights, and exceptional outcomes the more likely you are to venture further away from the monoculture that is slowly being created. Suddenly you can generate a uniqueness, a richness in ideas that everyone can only dream of… and what’s great is that this is pretty straightforward to get off the ground. It just requires a little awareness at the beginning (if only we had this all those years ago).

You know what makes poor soil? planting the same thing over and repeatedly. You might start off well but eventually your crops will be weak and sad. A cycle of decay has begun and the very thing that was providing life you have managed to destroy. Not taking care of the soil, not respecting the natural order of things turns out to be a bad thing - who would have thought?

We see the same with Gen A.I. (we’re really sweating this analogy huh?) when we all use the same inputs on Gen A.I. trained on the same data we begin to get pretty similar (and boring) results… and whilst Gen A.I. is currently trained on human data, in the foreseeable future a world exists where Gen AI begins to eat its own tail, teaching itself on data it created… who knows what the end of this loop looks like [ insert link ]

Breathe…. thankfully, you can mitigate this by making systems that keep space for fresh intel and input. There are ways to monitor this and test your model’s quality and stability. Always making sure that fresh, human made perspectives are coming in.

Be open-minded to change and train for the new world

‍Now obviously the industrial revolution didn’t happen overnight, it took decades for industries to adapt to these new ways of working and it took decades for the gaps that it made, between rich and poor, to become as evident as they are today. But just because something takes a long time doesn't mean we should ignore it for the first 30 years.

It requires an openness to the new potential, a willingness to train yourself, your peers to get those training laps in early.

We’ve been ploughing the land for some time now and whilst we may have gotten used to our ox you don't want to be stood next to it when the tractors roll in,

Be active and be open to new types of jobs and business models that will emerge.

‍
Things that earlier were impossible to do by hand, were filled with errors, or took an eternity when done by a machine, can now be done instantly and effortlessly. Exploring what this means for your business should be interesting and exciting.

‍

Summary‍

So what’s the moral of the story? What kernels of truth can we take away? Has the industrial revolution taught as anything?! Well with any new groundbreaking technological shifts it’s important to be cautious, think about the path you want to follow and don’t blindly accept the off the shelf tools and processes, these quick fixes may be the very thing that leads to your demise.

Begin to think about the knowledge your organisation holds, consider what might be possible if you were able to train a model on all your wins, all your best use cases and all your processes. Consider what your agency could do with a connected brain.

And whilst that might be a stretch for the imagination simply look at your output today, what tasks aren’t hugely strategic or creative but still important and a little repetitive? Your agency has all it needs to thrive and grow, just make sure to look after the soil.

‍