To retain agency over AI, we need to learn over and over again

Artificial intelligence
Image: Adobe Stock/putilov_denis

AI is a fixture in our lives, so how do we embrace and live with it, asks Antony Slumbers.

Artificial intelligence is at that stage where the hype cycle becomes irritating. Endlessly hearing about how AI will fix ‘everything’, and that you must do X or Y or Z or ‘risk falling behind’ starts to grate. To the point, I think, where many are getting turned off.

There is a limit to how much lecturing we can take.

Yet I think we need to take a few steps back, disconnect from the noise, and ask: what is actually going on here, and why should I care? The reason we should care is encapsulated in the example of exponential growth. If you take something very small and double it repeatedly, it looks for a while like not much is happening, but then the amounts get bigger and bigger. Slowly, then suddenly, everything changes. And this is what is happening with AI.

Take the graphics processing units (the computer chips which power most of AI) produced by Nvidia. These have increased in power 1,000 times in just eight years. Or the algorithmic models developed by the leading AI research labs. These have grown by an order of magnitude of 10 every year for the last 10 years. According to Jenson Huang, chief executive of Nvidia, they are running at “Moore’s Law squared”. And this looks likely to continue for at least the next two or three years.

So we are genuinely in a world that is seeing technology develop at unprecedented rates. And the hype merchants are correct in that. Where they are unhelpful is in telling you that tomorrow is today. Watch a demo and you can be enthralled, but a demo is in a controlled environment. In reality, many of the most exciting capabilities of AI are still at the brittle stage. Beautiful, but fragile. Ready for utopia rather than real life.

‘Real estate is a long-term game. So we need to be thinking many iterations of AI down the line and imagine what then might be possible, and what we’d like to be so.’

Antony Slumbers, Digital strategist

This might all be making you more confused – but don’t be. Given exponential trends, we are probably underhyping the near future. It is true that the price of intelligence is trending towards zero, and that the level of intelligence available to us is going the other way. As models get smarter, they are also getting much cheaper. A million tokens (a measure of words and parts of words) of GPT-4 cost $50 two years ago: today the equivalent costs $0.12. And this looks to be a solid trend, in the near term at least.

How, then, should we think about AI? Within a real estate context I think there are a few steps we need to take.

First, we need to understand the difference between predictive/analytical ‘classical’ AI, and its creative cousin, generative AI. They are good for different things: the former is perfect for using historical data to build predictions about the future; the latter is all about ‘generating’ new content (text, audio, video) based on a statistical model trained on, essentially, the entire internet. They each need different inputs, and their outputs are complementary, but not the same.

Predictive AI v generative AI

Second, given this knowledge, we need to think about how this might allow us to improve our existing businesses, but more importantly, what does this enable that simply could not be done previously? And what are the required inputs and outputs?

Third, how will these capabilities change how we can do business and the products/services we can offer? What skillsets will we need to make these happen?

And fourth, and in many ways most important, how will all of this change the nature of demand from our customers? What asset types will rise, or fall? How will form factors (space, amenities, structure) change? What will constitute ‘value’ in their eyes? How will the operating model of real estate change? Fundamentally, what real estate should we own?

Real estate is a long-term game. So we need to be thinking many iterations of AI down the line and imagine what then might be possible, and what we’d like to be so. The key, I think, is being very well informed. AI will permeate our lives entirely, and to retain agency over that we need to think and learn, over and over again.

Antony Slumbers is a digital strategist and product leader in the proptech space

Author: