Guest Lecture at Surrey University: What good is AI, if we can’t use AI for good?

5 key principles to use AI for good

Last week we had the pleasure of being invited to guest lecture at Surrey University for a mixture of Masters, PhDs and CPD students in Sustainability.

eco-shaper has had the privilege of a long, established and formal relationship with Surrey University; together we have successfully worked on projects for the UK government and the European Union. Surrey University’s Centre of Environment and Sustainability (CES) is an internationally acclaimed centre of excellence on sustainable development. Their research uses interdisciplinary approaches to the analysis of complex systems and from this they develop action-oriented, policy relevant responses to long-term environmental and social issues. The impact of their research and teaching is enhanced by their work in close partnership with diverse businesses, such as eco-shaper, on sustainable innovation and implementation issues in practice.

On this occasion, we were asked to provide the students with a business and technological viewpoint on carbon reporting. It was important for the students to understand carbon reporting from not only an academic/textbook point of view, but also to be able to apply their learning to the real world.

The lecture was based around 5 key principles when using technology for sustainability reporting:

1. Know your outcome: for most, it’s emission reductions; keeping global warming below 1.5ºc. The contradiction here is that technology is becoming the footprint multiplier. For example, training AI models puts a huge strain on processing power due to the amount of data needed to train the machines.
2. Explainable: When using technology for carbon reporting, one still needs to be able to explain the results. This relates to both being compliant with government standards but also thinking about what rules are baked into the algorithms. Challenges in sustainability frequently stem from carbon reporting not being explainable for corporations and leading to greenwashing claims.
3. Human Centric: Whilst technology can play a vital role in carbon reporting and reduction, it is key that sustainability remains human centric, as it’s an outcome of human based decisions. For corporations, sustainability and technological goals must be led by the business as relying too heavily on technology and particularly AI, can cause bias and inaccurate averages.
4. Transparency and controllability: Surrounding the topic of ‘ethical AI’, it’s important that the technology solutions we use for sustainability are transparent. Companies must stress to alleviate ethical concerns by protecting any customer or company data that is sensitive. While companies cannot remove the biases inherent to AI systems trained on large quantities of data, they can work to minimize adverse effects and not keep business decision separate. In other words, the technology should not exist in a black box.
5. Is AI a powerful hammer? The last point opened a debate around the topic of ‘what good is AI, if we cannot use AI for good?’ In summary, if technology is to be embedded in carbon reporting, it should be done in a resource measured way that is business led. It shouldn’t be generalized nor use large amounts of data that uneconomical and hard to trace.

We’d like to thank Surrey University, and particularly Dr Lirong Liu who works in Surrey University’s Centre of Environment and Sustainability department, for the continual and supportive relationship with eco-shaper, as well as allowing us to be a part of the integral education they’re providing to their students for our future.