UnEarth Magazine.

Words by Clarissa Bell, Art by Iona Maclaren

It’s my last semester of my undergraduate degree, and I won’t lie to you, reader, I’m feeling a tad nostalgic. This time two years ago – even typing that makes me feel old – I ended up writing a last-minute article for this magazine on corporate greenwashing. How original. My interests have evolved since then, delving more into how sustainability should interact with technology and social inclusion. Goodness knows my writing has also improved drastically since then. So, as I’m back on this little nostalgia trip, why not give one of my first articles for this magazine that proved foundational in my university experience a makeover? A glow-up, if you will. What’s more, I chose to re-write this article the day of the draft deadline upon finding out about OpenAI’s new tool: Sora. It’s about time I gave a few of my articles a 2024 update on the one topic that frustratingly remains an omnipresent Godzilla-scaled parasite in future digital sustainability politics: greenwashing. Any-who, memory lane is calling – and OpenAI won’t let us hang up.  

Greenwashing, according to an article by the United Nations, includes deliberate corporate tactics to mislead and misreport the public, and promote false environmental and sustainability solutions, progress, aims, standards, attributes etc that undermine efforts to reduce carbon emissions and effectively address climate change. As much as I admire this article’s example of highlighting the fashion industry and pointing us in the direction of how to spot it, in the era of rapid AI development, and a feverish hype around Silicon Valley-esque technological innovation that isn’t going to die down anytime soon, its advice is now outdated in 2024 despite being less than 2 years old. ‘Learn more’, ‘look for transparency and accountability’, ‘spend wisely’ and ‘considering a product’s lifestyle’ could’ve been considered a good first step to combat greenwashing in 2022. The question remains, though: can we use AI to detect greenwashing? 

Well…yes! An academic paper has been published on the use of AI and conceptualising greenwashing, and AI-detection appeared to be at the forefront of last year’s Financial Conduct Authority TechSprint competition. As of roughly 4 months ago, an AI Greenwashing Investigating engine ‘Athena’ from 15Rock is claimed to use real-time environmental datasets to expose false or inflated sustainability claims in the hope of paving future corporate accountability. Along the same line, GreenWatch, a research team and green claims detection tool based in University College Dublin, assesses green claims made by companies and provides data based on factors such as Greenhouse Gas Performance and Sustainability Claims. Similar tools have been made by EY and their ‘Greenwashing Compass’ prototype in October last year, not to mention PingAn in their Digital Economic Research Centre in China. None of these discussions should be new, though: in 2016, the Task Force for Climate-Related Financial Disclosures – whose prime task was making frameworks for public organisations regarding their performance on climate indicators – created ClimateBert for the same purposes, analysing over 800 companies.  

So, can AI be used to detect greenwashing? Well I’ll be darned, there’s a positive twist in one of my articles! I never thought I’d see the… hang on a second.  

With so many companies obviously using AI to detect greenwashing, the basic answer is: yes. AI can be used to this end. However, just like we need to be aware of potential greenwashing of company claims in their reports, I’m going to argue here that we’ll need to remain critically aware of the greenwashing analysis tools being reported by the same companies. This is because these were never meant to be products for us, anyone not inside financial companies or investing firms, to use. We’re probably not going to see a rise in public use of AI greenwashing detection until a much later date. Tools that release their findings to the public do so in non-lay terms, making it more work for anyone not already knee-deep in this topic. However, OpenAI have given us, and anyone who wants to use it, the opposite just a couple days ago by launching Sora. They’ve given us the tools to complicate things further.   

I’ve talked about the actual carbon and environmental impacts of these sorts of AI tools last year, specifically on OpenAI’s other AI darling, ChatGPT. However, Sora’s ability to take a picture or text input and make a photo-realistic minute of footage, even in infancy, is terrifyingly remarkable. Whilst I have to acknowledge that they’ve put a safety section on their main page, I know of no up-to-date detailed report (at time of writing) on the specifics of how they plan to engage “policymakers, educators and artists” in the future in a way that doesn’t involve implicit extraction of real-world scenarios and previous data. The previous data which is now under fire for greenwashing? 

Ergo, we have an AI-sustainability paradox on our hands. We, the public not usually privy to the use of this detection technology, are invited to be wary of what we consume and what claims companies are making, whilst having very little to no fast tool to detect it as quick as the companies’ investors. The companies can seemingly claim innocence and expose market competitors by creating a carbon-intensive technology to sell to investors to persuade them that their money is going somewhere sustainable. This is all whilst being able to create and sell to us another free-to-use carbon-intensive AI that has major implications for future sustainability credibility and disables our individual agency to… what was it again? ‘Learn more’? ‘Look for accountability’? ‘Spend wisely’?  

It would be disingenuous of me to end on a doom-and-gloom; we can’t afford to be hopeless when it comes to climate and sustainability action. Much of this action should be in the form of actual regulation, laws and nuanced environmental technology policies so everyone is able to make informed decisions. Social sustainability should be just as talked about as effective economic and environmental policy. Any-who, we can anticipate that digital literacy is going to become more crucial to our attitudes toward corporate sustainability than ever before, whether we all want it to be or not, and we need more guidance to keep up with these developments.  

Well, here we are, reader; the proverbial end of memory lane with the pot of unearthed AI-code at the end of the glitching rainbow. Here’s to looking forward, onward and upwards to the future.