From games to goals: Harnessing AI for business innovation and sustainability

DeepMind’s unbeatable AlphaGo has mastered one of the oldest boardgames in existence so well that its moves are considered ‘beautiful’[1]. In 2022 AlphaGo entered a five-game tournament in Soul where it beat Lee Sedol, the best Go player of the past decade[2].This AI not only learned how to play Go, it learned how to beat the creative intuition of one of the best human players of the game.

Whilst this is incredibly impressive and humbling, what does this mean for businesses and ESG? Effective matchmaking of AI technology with ESG technical expertise is a priority challenge. Let’s consider how AI can make the leap from boardgames into business operations:

• Carbon sequestration: The robust measurement and monitoring of agriculture, forestry and other land use (otherwise known as FLAG) and greenhouse gas removals (otherwise known as sequestration) is a challenge SLR is finding solutions for. This is a key strategic area in the context of Net-Zero, especially for organisations setting Science Based Targets[3]. New-ish technologies (such as remote sensing with predictive models using machine learning) are a source of hope for all of us looking for a non-resource intensive way to monitor carbon stocks and report removals. The challenge here is matching the technology with the technical standards[4], some of which are still in draft[5].

• Transport: Transport planning software isn’t a novel technology; logistics companies have used digital solutions to reap efficiency savings via digitally enhanced route planning for years. However, AI can up-the-game further. Predictive analytics is being used to level up-regression analysis as a means of anticipating changes in demand[6], this means organisations can better anticipate future routes and resource requirement. Generative urban AI is poised to transform urban design and transport infrastructure optimisation, in conjunction with being able to better anticipate population changes and travel preferences; this will help to solve a riddle that has historically stumped urban designers[7]. Finally, AI has been trialled in driverless cars, the current issues and ethics of which, almost go without saying.

• Energy: Back in 2018, DeepMind applied machine learning algorithms to 700 megawatts of wind power in the central United States to predict power output 36 hours ahead of actual generation[8]. Renewables still present a challenge to the power grid due to their intermittency and difficulty to plan for in real-time. VTT Technical Research Centre of Finland’s solution, VTT EnergyTeller is an example of AI being used to accurately forecast future energy needs and market developments today. In a country where a 1% error in wind power forecasting costs roughly €300,000/year in imbalance settlements[9], this is a true success story.

“We can only see a short distance ahead, but we can see plenty there that needs to be done”[10]. Alan Turing’s 1950 comment is apposite today. At this point, we have a fair understanding of the potential of different AI technologies. We know we’re about to see a disruption to how we do things (hopefully for the better) and we can point to some practical examples. Many of us who work within the built environment, industry, transport, or energy sectors don’t yet know how AI will impact our day-to-day work, or what AI means for the environmental and socioeconomic issues we face. We do know the key to success is matching-up the right people with the right technology.

We are beginning to understand the complex interplay and trade-offs that are inherent in our decision-making journey to towards Net-Zero. The kind of systems thinking possible through AI has the potential to be a significant game changer.

-----------------------------

References

[1] - The Sadness and Beauty of Watching Google's AI Play Go | WIRED: https://www.wired.com/2016/03/sadness-beauty-watching-googles-ai-play-go/

[2] - What we learned in Seoul with AlphaGo (blog.google): https://blog.google/technology/ai/what-we-learned-in-seoul-with-alphago/

[3] - Forests, Land and Agriculture - Science Based Targets: https://sciencebasedtargets.org/sectors/forest-land-and-agriculture#:~:text=Key%20requirements%20of%20the%20SBTi%20FLAG%20Guidance&text=Set%20long%2Dterm%20FLAG%20science,term%20FLAG%20science%2Dbased%20targets.

[4] - VM0032 Methodology for the Adoption of Sustainable Grasslands through Adjustment of Fire and Grazing, v1.0 - Verra: https://verra.org/methodologies/vm0032-methodology-for-the-adoption-of-sustainable-grasslands-through-adjustment-of-fire-and-grazing-v1-0/

Publications - IPCC-TFI (iges.or.jp): https://www.ipcc-nggip.iges.or.jp/public/gpglulucf/gpglulucf.html

[5] - Land Sector and Removals Guidance | GHG Protocol: https://ghgprotocol.org/land-sector-and-removals-guidance

[6] - Abouloifa, H., Bahaj, M. (2024). Artificial Intelligence in Supply Chain 4.0: Using Machine Learning in Demand Forecasting. In: Gherabi, N., Awad, A.I., Nayyar, A., Bahaj, M. (eds) Advances in Intelligent System and Smart Technologies. I2ST 2023. Lecture Notes in Networks and Systems, vol 826. Springer, Cham. https://doi.org/10.1007/978-3-031-47672-3_14

[7] - Generative Urban AI Is Here. Are Cities Ready? (forbes.com): https://www.forbes.com/sites/timothypapandreou/2024/02/18/generative-urban-ai-is-here-are-cities-ready/

[8] - Machine learning can boost the value of wind energy - Google DeepMind: https://deepmind.google/discover/blog/machine-learning-can-boost-the-value-of-wind-energy/

[9] - Energy Forecasting Software | EnergyTeller | VTT (vttresearch.com): https://www.vttresearch.com/en/ourservices/vtt-energyteller-energy-forecasting-service

[10] - A. M. TURING, I.—COMPUTING MACHINERY AND INTELLIGENCE, Mind, Volume LIX, Issue 236, October 1950, Pages 433–460, https://doi.org/10.1093/mind/LIX.236.433