AI and climate change: A growing communications challenge

This summer, Microsoft, the largest backer of ChatGPT developer OpenAI, revealed that its 2030 net zero “moonshoot” seems unlikely to succeed as a result of AI developments.

This was followed by Google also admitting that AI is threatening its environmental targets after revealing that datacentres, a key piece of AI infrastructure, have helped increase its greenhouse gas emissions by 48% since 2019, leaving “significant uncertainty” around reaching its target of net zero emissions by 2030.

The artificial intelligence boom has been felt universally. With AI already transforming aspects of our everyday lives from self-driving cars to medical diagnosis tools, it’s no surprise that the UK AI market is expected to grow from its current valuation of £16.8 billion to £801.6 billion by 2035.

AI has an enormous capacity to increase everyday efficiency, driving innovation in people’s personal lives – and at business level. According to the CBI, around one in six UK organisations have embraced at least one AI technology, with tools like data management software, language generation and machine learning creating significant benefits for organisations of all sizes such as greater productivity, optimised workflows and reduced human error.

However, pledges to heavily invest in AI products are coming up against pledges to reduce carbon emissions – and this backtracking on net zero commitments has the potential to cause serious reputation damage.

AI and climate targets

In January this year, the IEA issued its forecast for global energy usage and included, for the first time, projections for the electricity consumption associated with data centres and AI.

It estimated that this usage represented almost 2% of global energy demand in 2022 – and is predicted to double by 2026, which will make it equal roughly to the amount of electricity used by Japan.

This sharp uptick in the use and development of AI poses a serious threat to climate aspirations and companies’ sustainability strategies, especially those leading the charge in terms of innovation, such as Microsoft and Google.

As a result, big tech companies have become major purchasers of renewable energy and carbon offsets in order to try and keep up with their own climate goals (see a previous blog about the reputational risk of carbon offsetting here).

Amazon, for example, is the largest purchaser of renewable energy globally and Microsoft has now signed a record carbon credit deal with Occidental Petroleum in an effort to offset its increased emissions, which have risen by almost a third since 2020, mainly from the construction of AI data centres.

Other tactics have included heavy investment into nuclear power. After decades of stagnation in nuclear investment and development, this year has seen a rush of demand from companies like Google, Amazon, Meta and Microsoft, who are looking to nuclear to solve their net zero conundrum.

But the power use of a data centre is only part of the concern around carbon impact.

In Microsoft’s own emissions reporting, the company said its scope three (or indirect) emissions have been trending in the wrong direction, largely due to the building and sourcing of materials to construct the data centres themselves – something which cannot currently be done with renewable materials and power.

The good, the bad and the ugly

Despite the current rise of AI-related carbon emissions, prominent figures in the space are urging against panic.

Bill Gates has claimed that artificial intelligence will be more of a help than a hindrance in achieving climate goals, as the technology will eventually allow countries to use less energy by making existing tech and electricity grids more efficient.

Elsewhere, it has been argued that AI has the capacity to increase the spread of climate dis and misinformation, equipping ‘climate change deniers’ with new tools to swiftly and quickly disseminate misleading content. This includes AI generated synthetic media (‘deep fakes’) which have already been used in campaigns against renewable energy projects.

The risk of AI tools being misused for dis/misinformation campaigns requires real vigilance – see our blog on the topic here.

However, it’s not all bad. There are also AI models being developed that may help to counter misinformation, and designed to be used for learning specifically about climate. ClimateChat is an example of this, which is a question-answer chat tool that is being developed to help bridge the gap between complex scientific climate related information and digestible answers to big climate questions.

Such systems have the capacity to facilitate communication between experts, policymakers, and stakeholders, enabling more informed decision-making and promoting climate change mitigation and adaptation strategies.

Reputation Management

In any case, an increased use of AI spells increased energy usage and therefore a potentially bigger carbon footprint – at least in the short term as the technology develops.

This poses a serious risk for organisations’ climate pledges and the reputation implications of backtracking or failing to live up to expectations and make progress. To put this in a commercial context, a  recent KPMG poll of British consumers found that over half of those surveyed said they would stop interacting with a company if they were found to have misled consumers and investors over sustainability claims.

Around 30 countries have adopted laws enshrining net zero commitments since the Paris Agreement, and, as of last year, 94% of UK-based companies within the Forbes2000 have set net zero targets. With pressure mounting to adopt AI practices into working patterns, whether these targets will be met is being called into question – spelling danger in the court of public opinion.

Particularly for users of AI systems who also have standing sustainability commitments and ESG focused brand positioning, how these tools are powered and their potential carbon cost is going to be an increasingly central consideration.

Put simply, the challenge is balancing sustainability with technological progress.

With expectations around AI adoption rising, and scrutiny of climate credentials increasing in intensity, how these decisions are communicated to stakeholders will be fundamental – and getting it wrong could risk derailing reputation.

Download your free AI Reputation Risk Map