High Energy: Could Climate Concerns Slow the Dash to Embrace AI

With everyone from film producers to parliamentarians expressing concerns about the impact of artificial intelligence, regulation of the technology seems imminent. However, with governments still trying to determine what form that regulation should take, Commercial and Technology Partner James Teare explains why it’s essential for companies to put their own AI policies in place now to avoid leaving themselves open to challenges in the future.

Barely a day passes without another breathless boast about the potential of artificial intelligence (AI).

Within the last fortnight, we've had predictions regarding the technology's likely impact on conditions such as Parkinson's Disease (https://www.thetimes.co.uk/article/f8cce79c-4049-11ee-8b31-3c9c533abb75?shareToken=e83322a9acdcceb0c7f2b3f626bb5328).

In addition, the Deputy Prime Minister, Oliver Dowden, has suggested that AI could be even more transformative for the UK's fortunes than the Industrial Revolution (https://www.theguardian.com/technology/2023/aug/12/ai-could-have-bigger-impact-on-uk-than-industrial-revolution-says-dowden).

Now, I don't want to sound like a sceptic.

I'm actually among those people who believe that AI genuinely has the ability to make a positive difference to the way that we live and work.

I have spent several years working both for and with organisations - both large and small - looking to capitalise on its merits in one way or another.

With that experience, I am aware of the importance of adopting technology at the right time, when buying into it makes sense from a business point of view and is capable of delivering the desired results.

Nevertheless, even I've been somewhat taken aback by the haste with which some organisations are rushing to embrace the latest AI services.

One new study has, in fact, estimated that more than half of all investment made by SMEs based in London is on technology, including data analytics and AI (https://www.cityam.com/embargoed-00-01-london-smes-commit-over-half-annual-revenue-to-embrace-tech-growth/).

There are a couple of reasons why an advocate like myself believes a little caution may be in order.

It's not just the looming prospect of regulation, something which I've referred to in another article on this 'blog in the last few weeks (https://www.bexleybeaumont.com/indiv-feature?id=199).

However, whilst the regulatory landscape for AI is still evolving, it is possible able to identify some of the financial and environmental risks right now.

That is because the computing hardware which supports our reliance on the internet and aids the development of AI - much of which is housed in enormous data centres - is energy hungry. Very energy hungry.

The more sophisticated that it becomes, the more data and energy which it needs to consume.

Research by the International Energy Agency last year showed that the workload handled by the world's data centres has increased some 260 per cent in the last six years (https://www.iea.org/energy-system/buildings/data-centres-and-data-transmission-networks).

That has a knock-on effect in terms of greenhouse gas emissions. A study by Yale University concluded that operation of the Cloud, for instance, has a greater carbon footprint than the airline industry ((https://thereader.mitpress.mit.edu/the-staggering-ecological-impacts-of-computation-and-the-cloud/#:~:text=A%20single%20data%20center%20can,as%20much%20as%2025%20percent).

It is a problem which has been taxing decision-makers for some time. In 2020, the European Union was alarmed to discover that data centres would account for just over three per cent of all the continent's electricity by the end of this decade (https://digital-strategy.ec.europa.eu/en/library/energy-efficient-cloud-computing-technologies-and-policies-eco-friendly-cloud-market).

Data centres don't only affect the environment in terms of the power which they need. All of the minerals, for instance, which go into building, equipping and maintaining them need to be produced somewhere. Those processes have their own specific impacts.

Such matters are now seen as a "priority on the EU political agenda" but should not be discounted by businesses as being beyond their immediate concern, especially when companies are now obliged to weigh up their responsibilities from an Environmental, Social and Governance (ESG) perspective.

Sustainability policies are far from a fringe element of a firm's activities.

Earlier this year, the UK Treasury described how just under half of the £10 trillion worth of assets managed by the UK financial services industry in 2021 "had integrated ESG into the investment process" (https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/1147458/ESG_Ratings_Consultation_.pdf).

In late July, the EU also introduced new rules requiring companies to their impact on the environment (https://finance.ec.europa.eu/news/commission-adopts-european-sustainability-reporting-standards-2023-07-31_en).

Contributing to the energy consumption of data centres is one part of those disclosures, albeit something which might not feature among the main considerations when the decision to approve spending on AI is being taken.

Such thinking is arguably of particular relevance for UK companies, due to the fact that this country ranks third among all global territories in terms of its number of data centres (https://www.statista.com/statistics/1228433/data-centers-worldwide-by-country/#:~:text=Number%20of%20data%20centers%20worldwide%202022%2C%20by%20country&text=As%20of%20January%202022%2C%202%2C701,456%2C%20while%20China%20recorded%20443.).

Many organisations now rightly regard a robust ESG policy as critical in facilitating growth, managing costs and optimising investment opportunities.

If AI is to part of a business strategy into the future, it surely makes perfect sense, therefore, to take time to determine how it is to be incorporated.

Should firms not build something which is sustainable from the start, there is always the possibility that they may have to strip it out at some later point.

If AI has become integral to their systems by that point, the hit to their operations and their balance sheets might be substantial.

When firms think of future-proofing, it is tempting to think only of the benefits. We can have no accuracy about whether regulation and the policing of ESG credentials will really bite and, if so, what the consequences will be.

To discuss any of the above further, please feel free to contact James Teare: jamesteare@bexleybeaumont.com  |  07709 733459