My colleague César recently wrote about how Power Apps and generative AI can help businesses achieve sustainability goals and reduce environmental impact. With the pace of AI change and adoption accelerating I wanted to expand on this a bit further.

A short while ago I had the opportunity to attend the GenAI Summit, an event that brought our clients together with experts and enthusiasts of generative AI from the DACH (Germany, Austria and Switzerland) region. The summit was hosted on a mountaintop called the “Patscherkofel” in Tyrol, Austria, where participants could explore the scenic landscape and interact with each other. It was an amazing experience.

As I was taking the cable car up to the summit, I couldn't help but think of how this mode of transportation was analogous to the journey of generative AI. For some people, being in a cable car can be a bit frightening, as they feel helpless and not in control of their fate. They may wonder what would happen if something went wrong, or if the cable snapped. Similarly, many people have fears and doubts about generative AI, as they see it as a black box that can produce unpredictable and potentially harmful outcomes. In a recent AI Readiness Report, only 48% percent of respondents said that they completely trust the results of AI. They may question the ethics and reliability of this technology, or its impact on human creativity and agency.

But just like the cable car, generative AI is not a mysterious force that operates on its own. We spoke to an AI executive from a multinational oil & gas organization, and they echoed the sentiment, “People often think AI is magic and it's not. It often doesn't work, so we expect most of our proof of technology or proof of concept projects not to go through to full scale up because that's just not how it works. You do it to test it out and then the ones which are working well you continue on with. You need to try lots of different things to see what works and to see if they'll work in your context.”

AI is a tool that has been built by humans, is operated by humans, and is maintained by humans. And it only takes us from one point to another, it does not determine our final destination or our actions along the way. We still need to walk to the station and continue to walk from where we arrive. We still need to set the goals, provide the data, evaluate the results, and apply them in a meaningful and responsible way. Generative AI is not a threat to our autonomy or our values, but an enabler of our innovation and our sustainability.

This last point was especially evident at the summit, where I met with our clients to discuss some of the exciting and inspiring applications of generative AI for sustainability. Below examples offer a glimpse into the future of how we can use this technology to reduce our environmental impact and achieve our sustainability goals.

- ESG report analysis with generative AI: This is a solution that uses natural language processing to help businesses analyze their performance on environmental, social, and governance (ESG) criteria, and compare it with their competitors and industry benchmarks. The solution can automatically extract relevant information from various sources, such as annual reports, sustainability reports, press releases, and news articles, and generate summaries and insights on how the business is doing on different ESG dimensions. This can help businesses identify their strengths and weaknesses and prioritize their actions for improvement

- Practical sustainability actions powered by Microsoft Sustainability Manager (MSM): This is a solution that uses reinforcement learning and natural language generation to help businesses create project proposals for sustainability improvement. The solution can learn from the data and feedback in MSM and suggest possible projects that can help reduce the carbon footprint, water consumption, waste generation, or energy usage of the business. The solution can also generate a draft proposal for each project, including the goal, scope, budget, timeline, and expected outcomes. This can help businesses save time and resources and accelerate their sustainability transformation. This can augment a service like Avanade Sustainability Quick Start which helps companies track emissions across scopes 1, 2 and 3 and gain insights into actions to reduce emissions using MSM.

- Sustainability chatbot (custom copilot) in MSM: This is a solution that uses natural language understanding and generation to help businesses access relevant and reliable information on sustainability. The solution can act as a personal sustainability advisor that can understand the intent and context of the user's queries, and provide answers accordingly from curated sources, such as academic journals, reports, websites, or podcasts. The solution can also engage in a natural and conversational way with the user, and offer suggestions, tips, or feedback on their sustainability efforts. This can help businesses increase their awareness and knowledge and foster a culture of sustainability within their organization.

- CSRD reporting tool on Power Platform: This is a solution that uses natural language generation and document automation to help businesses comply with the new Corporate Sustainability Reporting Directive (CSRD) of the European Union. The solution can automatically fill 85% of the content for the CSRD report, based on the data and information provided by the business, helping to reduce the burden and complexity of reporting, and enhance the transparency and credibility of their sustainability disclosures.

These are just some of the examples of how generative AI can help businesses with sustainability. There are many more possibilities and opportunities that can be explored and realized with this technology, however, while generative AI can offer great benefits for sustainability, it also poses some challenges and risks that need to be addressed. One of the main concerns is the environmental impact of generative AI itself, as this technology requires a huge amount of computational power and resources to train and run. Technology is a hidden cause of carbon emissions and its for this reason that introduced Avanade Green IT Enabler, a service aimed at helping organizations automate monitoring and take practical actions to reduce carbon impacts of their information technology (IT) infrastructure and operations. Depending on the size of the Large Language Model and the type of energy used, training an LLM can consume between 25 and 500 metric tonnes of carbon dioxide, the equivalent to over a million miles driven by an average gasoline-powered car. Therefore, generative AI can potentially undermine its own sustainability goals, if not used in a responsible and efficient way.

-Use renewable energy sources or carbon offsets to power the generative AI systems

-Optimize the data and models to reduce the size and complexity of the generative AI systems

-Evaluate the necessity and value of generative AI for a specific purpose or problem, and explore alternative or complementary solutions

- Ensure the quality and reliability of the data and models, and avoiding biases or errors that can harm the environment or society

- Provide transparency and accountability for the generative AI systems, and disclose their methods and outcomes

- Engage with diverse stakeholders and experts to assess and address the potential impacts and risks of generative AI

By following these guidelines, we can harness the power and potential of generative AI for sustainability, without compromising our own environmental and social responsibilities. Generative AI can be a powerful ally for sustainability, but only if we use it wisely and ethically.

To learn more about the role that technology plays when it comes to sustainability, visit smart sustainability actions that Avanade offers.