'Companies need to ensure sustainable AI is more than just a bullet point'. What steps are needed?

Currently, over 54% of marketers use AI to achieve faster content generation and campaign rollout. Other popular uses of AI by marketers include creating content that resonated with the emotions of the user (43%), automating the creation of segments based on the probability of conversation (39%) and running experiments at scale (39%).

However, little do marketers know, the use of AI leaves an environmental impact, something that tech giant Google brought to light in its annual environmental report where it said its greenhouse gas emissions totalled 14.3 million tonnes of carbon dioxide throughout 2023. This is 13% higher than 2022, and 48% higher than 2019.

The increase in emissions, according to Google, is primarily due to increases in data centre energy consumption and supply chain emissions required to integrate AI into its products. With tech giants facing these issues, how can marketers then use AI sustainably?

Don't miss: GroupM and SeenThis partnership dodge over 230 tonnes of carbon emissions

June Cheung, head of JAPAC at Scope3, said that marketers can first start by understanding gen AI's implications and how it's not necessarily a more sustainable option.

"To responsibly harness this technology at scale, marketers need to know the climate impact of their work, full stop. That starts with working with partners that can help you measure and understand the emissions from how exactly AI is being used," said Cheung.

With that information, marketers can align closely with the sustainability and executive leaders internally on how the use of gen AI will impact business outcomes and ESG goals, explained Cheung.

"Businesses, including Lego, are already tying employee KPIs both emission reduction and business outcomes; they’re intrinsically linked. In that way, being aware of the climate impact of your work and keeping that number in check is now imperative," said Cheung.

Hand in hand with keeping numbers in check is transparency, explained Cheung, who said "CEO’s don’t want to be in shock that their marketing department has blown the company's sustainability targets out of the water and have difficult conversations with the Board and market."

At the same time, marketers need to lobby their leadership to make carbon footprint reporting an integral part of the company culture and mission statement, said Don Anderson, CEO of Kaddadle Consultancy.

They need to establish sustainable AI policies that clearly address the 'when, where and why' application of AI in marketing processes.

"If a company is sincerely concerned about its carbon footprint, then their employees, particularly its marketers and comms teams who are most likely to be in the front line of AI tool deployment, need a set of questions in front of them to guide their decision-making when it comes to employing AI," said Anderson. He added that marketing leads should help define and reinforce the frequency and degree of AI tools deployment and individual reliance in creative production.

Keeping marketers accountable

To keep themselves accountable, marketers can consider investing in responsible AI, said Mary Catherine Fixel, head of sustainability studio at Accenture Song.

“Many of the first and easiest use cases to scale AI are in marketing, so it’s critical we invest in figuring out responsible AI – what that means and how to deliver on it, sustainably," said Fixel.

"We envision a future where every marketer measures and manages the emissions from and efficiency of their marketing, including the application of gen AI," she added.

In addition, CMOs and their teams are starting to think about how they can decarbonise marketing to deliver better performance, said Fixel. However, a real systemic shift in the industry is needed to scale these ambitions.

AI scientist Milind suggests measuring efforts by tracking energy consumption through the use of tools provided by cloud platforms to track the energy used during AI training and inference. He also added that utilising Google's environmental impact calculator to estimate the carbon footprint of AI activities and comparing AI usage with industry benchmarks to identify areas for improvement are ways companies can measure their carbon footprint and be accountable.

On top of that, marketing agencies can look into developing policies for sustainable AI practices and communicating these policies to all stakeholders, added Milind.

Agencies can also consider including AI-related carbon footprint in overall sustainable reports and work with industry groups to develop standardised methods for measuring and reporting the carbon footprints of AI activities.

Anderson agrees, adding that senior leadership and internal teams or committees should design specific guidelines or criteria for measuring and tracking the impact of any marketing initiatives involving technologies that impact the organisation's carbon footprint in consultation with known and respected environmental or non-profit bodies committed to establishing global standards and approaches.

"These internal bodies would establish clear guidelines that are regularly communicated top to bottom within the organisation, and make it their corporate mission to ensure that their customer base and partners are aware of their ongoing initiatives through key communications, including official environmental impact reports and audits developed in collaboration with third-parties," said Anderson.

"In turn, this information can be leveraged in B2B/B2C newsletters, PR and media briefings, promotional videos and investor reporting communications to validate the company’s efforts in this space," added Anderson.

Managing carbon emission goals

However, most companies are still struggling to maintain pace with the effective uptake and implementation of AI tools and technologies in their own companies and among their employees. At the same time, many CEOs still don’t have answers on how AI investments will result in real returns for these businesses.

"I’m not convinced that the majority of companies are prepared to manage similar carbon emission issues in this new age of artificial intelligence. I doubt very much that companies are making this a priority dialogue, or even know where to start when it comes to establishing sustainable AI policies for its marketers or partner vendors to follow," said Anderson.

"We’re simply not there in making this a critical , and I’m afraid the aggressive stance of industry in trying to push AI technologies adoption without full consideration of the environmental impacts, will blow up in our collective face," added Anderson.

For Anderson, managing carbon emission goals comes down to education, adding that:

Companies need to ensure that ‘Sustainable AI’ is more than just a bullet point within internal AI ethics and management training and education programmes.

It's also the responsibility of external education providers such as Singapore Management University (SMU), which have developed bespoke certifications around gen AI ethics, to reinforce the potential environmental impacts of AI implementation, said Anderson.

Cheung agrees, saying that she thinks of the emissions associated with AI as coffee cups. 

"Prior to knowledge and education we didn’t know how long a plastic cup would last in the environment; we didn’t give a second thought to our takeaway coffee. But now that we know, consumers have the choice of bringing their own cup, with many of us taking that option," explained Cheung. "This means four things - measurement, transparency, evaluation and action."

According to Cheung, companies need to know the comparative environmental cost of AI tools. For example, companies such as Google and ChatGPT should transparently display the emissions cost of a query and give the market the option.

With accurate and granular measurement and increased transparency into the impact, employees and companies can evaluate how they leverage AI based on the use case and trade-offs, said Cheung.

"If we know ChatGPT generally uses more energy and as a result more emissions than a standard Google query, is using GPT for a basic piece of information necessary? Does it make more sense to use AI in the production process if it means flying 10 fewer crew to a remote location to film?," said Cheung.

Having the right data and information gives the industry what it needs to take the best course of action to eliminate unnecessary use of AI and to transparently document the impact of its use to ensure it aligns with both business and ESG outcomes, she added. 

Gear up for the most captivating PR conference of the year! #PRAsia #Singapore &  #Malaysia returns, bringing you the tools and insights you need to dominate in this dynamic environment. Join over 100 industry leaders and experts for an action-packed program filled with insightful discussions and valuable takeaways.

Related articles: 
Clean Creatives calls out more B Corp agencies for fossil fuel deals 
Hitachi's new sustainable ad project aims to cut carbon emissions by 25% 
Do brands need to be transparent with consumers when using AI content?