Is AI an artificial investment in a greener future?

Huw Dixon, senior partner at specialist building consultancy HartDixon, is a great believer in embracing technology but urges caution against rushing to entrust computers with conservationism.

Given recent media headlines, artificial intelligence’s current inability to gain sentience is perhaps a deliberate act of self-preservation.

Had it already got its micro-processing chips around registering experiences and emotions, AI would undoubtedly be feeling bruised by suggestions from experts – including the heads of OpenAI and Google Deepmind – that, in the extreme, it could lead to the extinction of humanity. 

Indeed, you’ve got to have incredibly thick skin (or be a computer) to simply brush off being ranked alongside other societal-scale risks such as pandemics and nuclear war, and linked to disaster scenarios that include the creation of devastating chemical weapons, propping up rogue regimes and enfeebling humankind.

AI does, of course, have more than its fair share of advocates, who have been quick to say such fears are overblown and proffer plentiful pros to counter the cons currently consuming newspaper column inches. 

There are also those who, for the time being at least, are happy to occupy the middle ground and, when it comes to the application and utility of AI by businesses impacting and within property and construction, I’d argue the cement is yet to set on a consensus.

That is certainly true at HartDixon. In the workplace, we see its positive impact – in the guise of automated intelligence – in support of specific business processes and its wider development is being tracked with interest. Having a tool that captures minutes and intuitively presents meeting notes may not sound very sci-fi or cutting edge, but it is already proving a genuine time saver and I can see some clear benefits that AI can offer to the industry as a whole in the not-too-distant future. 

Fan or not, the impact of AI will be felt by the property and construction markets. When a company such as Octopus Energy reveals that 34 per cent of its customer email enquiries are now being answered by AI – just three months after its adoption – then it’s clear changes to office property markets and occupier fit outs are coming. Computers, after all, do not require the same desk space as their human counterparts. Conversely, the increased use of AI will drive demand for data centre facilities, increasing investment and development in a sector already very familiar to HartDixon.

I genuinely see both opportunity and uncertainty arising from the use of digital technology.   

Chief among the advantages must be the potential to bring greater sustainability to construction and property. From identifying and eliminating conflicts between structural, MEP and architectural design on new construction schemes to optimising planned maintenance programmes and energy use in operational buildings, AI would appear to afford a more environmentally-sound tomorrow.  Why wouldn’t enhanced efficiency and intuitive management of services reduce carbon footprints?  

For a company close to achieving its target of net zero, the prospect of further reducing our carbon emissions, and those of our clients, is extremely tantalising.

However, just as Professor Yoshua Bengio – who is hailed as one of the ‘godfathers’ of AI – recently conceded to wishing he had prioritised safety over usefulness, businesses need to be cautious about entrusting conservationism to computers.

The computation associated with AI, big data and machine learning comes with significant environmental costs. 

According to research by the Massachusetts Institute of Technology, the digital cloud now has a larger carbon footprint than all the airplanes flying above the literal clouds; a single data centre can consume an amount of electricity equivalent to 50,000 homes; and training one AI model can produce nearly five times the lifetime carbon dioxide emissions of an average American car or the carbon footprint of running 100 homes for a year. 

At first glance, that is a significant trade off to introducing workplace efficiencies – whatever your field – and one largely overlooked in favour of the more headline-grabbing narratives of people being replaced by machines and robot minds running amok. 

That is not to say that I believe AI does not have a role in, or is incapable of, helping to check climate change, but that the view from my desk at HartDixon is that there seems to be some important questions in need of answers.

Ironically, I would suggest one reason for that is that there is not yet the data available to crunch and produce a substantive and compelling case for or against AI in respect of sustainability. I’d love to know, for example, whether the eventual savings delivered by AI-planned eco-friendly, efficient buildings cover the upfront energy investment needed to train a tailored AI model. 

Until someone – or something devoid of any bias in its system – can give me a better-informed answer, I’ll be keeping an open mind.