Artificial intelligence — a vehicle or an obstacle on our path to a sustainable future?
Many people and organizations across the globe work towards a sustainable future. Artificial intelligence (AI) can be a powerful tool in their efforts.
AI technologies, which aim to imitate the natural intelligence of humans and animals, have made enormous progress in the past decade. AI has brought us a whole range of futuristic tools — from self-driving cars to virtual assistants, smart home appliances and self-learning robots — and enabled us to process large and complex data with unprecedented speed and accuracy. With AI advancing so rapidly and its application becoming more ubiquitous across society every day, we should carefully consider how to use it to ensure a sustainable future.
AI as a key asset for sustainable transformation
Scientists in many disciplines, whether they work in medicine, social science, engineering, climate or biodiversity research, have much to gain from using AI. Computer vision models can explore large microscopy datasets, enable efficient monitoring of wildlife and vegetation and can be used to automatically look through the steady stream of images generated by the large fleet of climate-monitoring satellites. Natural language processing models can run through millions of text pages, connecting and summarizing information to guide therapy development, environmental research and policy decisions. Other AI models can operate unmanned vehicles and sensors and analyse the data generated by them. AI thus enables scientists to collect and process high-quality data on an unprecedented scale, speeding up the pace at which they generate new sustainability-related knowledge.
AI can also be a powerful tool for public authorities. It can monitor water and air quality and support the response planning for natural disasters and pandemics. Authorities can also use AI to detect misinformation and operate chatbots and newsfeeds that inform general citizens and professionals alike.
For the private sector, AI can help steer energy grids according to actual demand, provide quality control and predictive maintenance as well as optimize logistics and factory operations to minimize costs and waste of resources. AI can also control industrial robots, autonomous vehicles and smart devices that increase efficiency, safety and convenience.
For ordinary citizens AI can do much as well. It can provide targeted educational programs, translate information into any language and power assistant services and devices for those that cannot cope on their own, giving them better access and more independence. If trained right, AI systems can also be less biased than humans reducing discrimination. Not least important, AI systems can automate or speed up tedious everyday tasks, freeing up time for social and cultural engagements.
These are just some examples. Like computation and electricity, AI can be used in almost infinite ways to bring us closer to the sustainable development goals.
AI as an obstacle for a sustainable future
However, AI also comes with flipsides, that could offset these benefits. AI can internalize, propagate and potentially even amplify existing biases, if trained with flawed data and these issues are exacerbated by lack of transparency. AI could widen the gap between high-tech countries and those that lack the necessary infrastructure to operate AI systems at scale and give unprecedented power to “gate-keeper” companies, if we do not develop strategies for ensuring equal access to AI technology around the globe. AI has been and will again be used to invade privacy and abuse human rights, if such applications of AI are not explicitly banned through international treaties and national laws, which are strictly enforced. AI will cause the disappearance of many types of jobs resulting in increased poverty and social upheaval if we cannot support the transition of large groups of people into different occupations or completely transform the way people obtain their income. AI development and operation requires massive amounts of energy, which will make it harder to minimize climate change unless both code and hardware become more energy-efficient, or energy is sourced from renewables on a much larger scale. The production of the computational hardware and the development of AI systems itself is also associated with a whole range of environmental and ethical issues — rare metal mining, electronic waste, inequality in the tech industry, production in low-income countries to name a few — which need to be addressed. And then there is the potential risk that future general AI systems — not the specialized AI we have today — might become too advanced to control and possibly even a threat, with even experts disagreeing whether this is a real possibility or simple fearmongering.
The path towards sustainable AI
The negative effects of AI need to be minimized so that they do not offset the enormous benefits we could reap from this powerful technology. If we get it right, AI could indeed help us transform our societies for the better and contribute to a more sustainable future. To achieve this, we need to from a broad coalition across research fields, sectors and nations with sustainability experts, AI developers and regulators working hand in hand to ensure this technology is used for good. Together, they can develop better AI technologies, accurate tools to measure their impact and strategies and regulations to avoid negative effects. And even ordinary citizens, from children to seniors, need to learn more about this technology, so that they can decide how to use AI in their own lives and take part in a fact-founded public discourse about how we should employ this technology and how we should not.
How can you learn more about AI and its relation with sustainability?
If you know more about AI technology and its opportunities and drawbacks when it comes to sustainability it allows you to make your own informed choices regarding its use. Here are some suggestions and resources:
1. Read articles about AI in your favorite newspaper and magazine. Then read some from a source that typically provides views that oppose yours.
2. Watch short film clips or documentaries about AI technology, its applications and its relation with sustainability. You can, for example, re-watch events from Lund University on AI Lund TV. You can also find many excellent videos on the pages of general or educational television stations (in Sweden we have UR Play, for example) or youtube (we have collected some on the playlists curated by our research group).
3. Take a free online course. For example, one of Lund University’s AI- or sustainability-related MOOCs on coursera (https://www.coursera.org/lunduniversity) or the Elements of AI course. During the pandemic, many countries have also produced excellent digital teaching materials for home-schooled students which often are freely accessible.
4. Check out your local science museum — many now have exhibitions on AI and/or sustainability. Many international science museums also offer virtual tours.
5. Read an in-depth analysis on AI and its relation with the UN’s Sustainable Development Goals. For example, this one.
