Guiding Questions for Implementing Artificial Intelligence as an Enabler for Sustainability

The potential of Artificial Intelligence (AI) as an enabler for sustainability is increasingly evident, but so are worries pertaining to data biases, data privacy, the environmental footprint and rebound effects. So, how to bridge the gaps between economic feasibility, ethical considerations, and environmental standards? And how can small and medium-sized enterprises (SMEs) make the most of it?

Artificial intelligence (AI) is not only one of the key technologies for the digitalisation transformation, but also a lever in the transition toward more sustainability. In the AI breakfast series organised by the Competence Centre eStandards, key application areas such as clean energy, circular economy or sustainable production were discussed from the perspective of small and medium-sized enterprises (SMEs). Due to regulatory guidelines, funding programs and competitive pressure, AI applications have reached a level of importance that can be hardly ignored by SMEs. But unlike large enterprises, these companies have generally less resources to integrate AI applications into their business, especially considering their high level of sophistication and complexity. The existing lack of generally-accepted frameworks and poor communication of practical examples only aggravates the situation.

How profit and a positive socio-environmental effect can go hand in hand

Despite the challenges, pioneers are successfully applying AI not only to generate profit and boost growth, but also in favour of the environment. The AI-powered software reduces the number of returns from online apparel sales by measuring the size of customers using their own smartphones. Smarter Sorting’s AI-based platform provides a sustainable solution to deal with unsellable or returned products by recommending sustainable and cost-effective solutions for disposal. Targeting farmers, Peat’s Plantix app provides smartphone-based AI advice on plant disease and the proper treatment. This results in reduced and optimized use of pesticides as well as the improvement of food security.

Why ethical considerations matter

Despite its benefits, AI can also induce several risks, especially when it comes to sustainability and ethical considerations. The possibility of carrying or replicating discriminatory practices, triggering overconsumption or promoting unsustainable production and consumption patterns, are a few of such risks. A recently published study found that 65% of 100 top firms from very different sectors were not aware how decisions or predictions based on AI are made exactly. This shows that even though businesses are well informed about the potential that this technology provides, there is still a lack of a structured and comprehensive way to deal with the inherent responsibilities that the application of AI demands. That is why a positive social and environmental effect of AI applications should not only be desired, but also considered upfront. For this purpose, the European Commission has introduced a set of rules and actions aiming to strengthen the trust in AI technologies. Instead of merely reacting to regulatory demands, there is much room for companies to proactively shape their application of AI technologies. Well-thought and coordinated plans ensure that the implementation of AI solutions is embedded in a proper framework to maximise its benefit for society and the environment.

If you are an SME who is exploring the idea of implementing AI solutions into your business, check out our list of guiding questions. This short questionnaire is designed to support you in uniting economic gains with positive socio-environmental results.

Key questions: Environmental sustainability

Cost-benefit considerations:
Is a computationally-intensive AI solution necessary for a given case or is a “classic” digital solution a better choice? Does the ecological benefit outweigh the negative impact caused by the energy-intensive training of the AI?

Choice of service provider:
Which hardware, data centres, and cloud or AI providers use green energy and are committed to sustainability? 

Pre-selection of datasets:
How to reduce the complexity of essential datasets for training the AI through pre-sorting in order to ensure less computational power, thus less energy consumption?

Key questions: Ethical considerations

Autonomy & control:
Is self-determined, effective use of AI feasible?

Does the AI solution treat all affected parties fairly?

Are the functioning and decisions of the AI application easily and clearly comprehensible?

Is the AI application reliable and robust?

Is the AI solution sufficiently protected against attacks, accidents, and errors?

Data protection:
Does the AI application protect privacy and other sensitive information?

We are keen to collaborate with you in finding solutions that align economic feasibility, ethical considerations and environmental impact in the implementation of all AI solutions. Do you share the same vision? Then, reach out to us!

For further questions, please reach out to Arne von Hofe.

Photo by Alina Grubnyak on Unsplash