AI In Sustainability Panel - Sustainability LIVE London 2024

December 03, 2024

undefined mins

Executives from Kyndryl, SAP, Google and IFS featured on a panel at Sustainability LIVE London that delved into the potential of AI for sustainability

At Sustainability LIVE London Global Summit, Faith Taylor, SVP of Sustainability, ESG & Impact at Kyndryl, Daniel Schmid, Chief Sustainability Officer at SAP SE, Adam Elman, Head of Sustainability for EMEA at Google and Sophie Graham, Chief Sustainability Officer at IFS featured on our ‘AI in Sustainability’ panel that explored how AI can drive a sustainable future when used responsibly and ethically.

Faith Taylor opens the discussion by highlighting Kyndryl’s use of AI in its 230 data centres, particularly for machine learning and data analytics.

“We've been implementing AI for many years. As you know, when you look at data centres, AI and machine learning has been used.

"And so that's the first place that many companies are utilising AI, but then they're using it also in terms of migrating to the cloud or also doing data analytics and consolidating information with regards to AI.”

Faith emphasises the importance of decarbonising AI to address its energy demands, comparing it to earlier concerns raised during blockchain adoption.

Daniel Schmid brings a broader lens to the conversation, discussing SAP’s integration of AI in ESG reporting.

He details how their control tower solution streamlines data collection for compliance, highlighting how “the AI generator reporting tool is embedded and brings down the manual efforts by up to 80% allowing time to bring in the business value.“

AI in Sustainability Panel at Sustainability LIVE London Global Summit 2024

Addressing AI’s challenges: Energy use and ethics

The discussion also addresses the significant energy demands associated with AI.

Daniel proposes innovative ways to mitigate this, such as synchronising data processing with periods of high renewable energy availability.

Adam Elman echoes these concerns, while noting advancements like carbon-intelligent load shifting and more efficient AI hardware.

“Our latest tensor processing unit - which is the key processor for AI - is 67% more efficient than TPU five equivalent. And that was more efficient than the model before.”, he shares, underlining Google’s commitment to sustainability in AI.

Another challenge discussed is the ethical implications of AI use, including bias in data sets.

Faith stresses the need for inclusivity: “AI is built on the data that we put into it. If we're only putting in data primarily, [...] it will be biased because it's it's built on our data.”

Sophie Graham adds that AI's design should involve diverse teams to better serve global communities.

The AI in Sustainability Panel highlights the need to integrate ethical practices when harnessing AI

Collaboration as the key to progress

All panellists agree on the importance of collaboration between policymakers, technologists and sustainability experts.

Daniel details SAP's ethical approach to AI, including adherence to UNESCO principles and the use of external ethics panels.

The conversation also highlights the responsibility of industry leaders to support equitable AI access worldwide.

Adam comments on the cloud's potential to democratise AI resources, while Faith points out the disparity in resources between developed and developing countries.

“Responsibility to me has a huge ethical component. [...] We need to think about it holistically - how do we include those communities and those countries that don't have access to have access in an equitable manner?” she emphasises.

Later on, Faith reflects on the inclusivity of AI development, stating how “a key component for solving this is how do you make sure the data is inclusive across the categories so that it does not have a bias.

"We're building AI every day so that's a huge opportunity in my mind.

"And we're still learning, this is not done. This is where we're in it right now.”

AI in Sustainability Panel - Sustainability LIVE London Global Summit 2024