There is no escaping the recent boom in AI products and tools entering the marketplace. It feels like every major online tool, app, and business is suddenly releasing an AI addon or full-blown product. Alongside the incredibly important questions surrounding ethics, and AI’s effects on humanity as a whole, is another important question. What effect is AI having on digital sustainability, and what continued effects may it have in the future?
The power of AI
There is one thing for certain. AI is powerful. It has the potential to both positively and negatively affect sustainability across a wide spectrum of industries. My focus is on digital sustainability. So I won’t delve much into other aspects here. The conversation is just too expansive for one post.
Let’s start with the potential positives.
How can AI have a potentially positive effect on digital sustainability?
AI can be used as a tool, just like any other tool we use at the moment.
Data centre efficiency
AI could be used to optimise the energy usage of data centres. Through the measurement of data, AI could be used to run different scenarios that could help to reduce energy consumption in the expansive, power-hungry buildings we all host our digital lives within. It could help to empower engineers to come up with new solutions for more energy-efficient server structure, material use and waste management.
E-commerce and efficiencies
AI could, and arguably has already, help delivery companies save on fuel. Providing better routes that are more fuel efficient with fewer miles travelled for the same number of deliveries. While not directly digital, these deliveries can be a result of e-commerce sales, powered by our digital services.
Search and speed
AI could help humans to achieve tasks more quickly. Take a common practice on the web now. Searching. We all use search engines. With AI integrated into these existing tools, it could potentially provide the answers we require quicker, and in an easier-to-understand written format. This does of course have huge ramifications for website owners, and how content is written and accessed by AI, but, it could have its benefits.
How can AI have a potentially negative effect on digital sustainability?
Data and power use
The trouble with AI is that it takes an incredible amount of data and computing power to train. With OpenAI suggesting a trend that sees an increase by roughly a factor of 10 each year
in the amount of compute power needed to train AI models, the problem of power use is only going to get worse. Generating power without fossil fuels is one of the many challenges of the climate crisis. With more and more power being used up by data centres, and AI needing increasingly more of it, will we see even more energy poverty in the world?
Data and ownership
The other big problem with AI is the data that it is trained on, learns from and uses to produce results. There have already been major issues and outcries from many different people around the rightful copyright of AI-generated content if they have been trained on someone else’s work. Certain tools, like Adobe, have released image-generating AI which is trained by the mass of works created via Adobe’s tools used by real people. They have added a setting that allows people to opt out of their work being used for AI training, but I worry that this has not been made clear enough. This is just one of many examples of a similar nature.
E-waste
The world already has major problems with e-waste. Huge swathes of products are thrown away every year, replaced by the next slightly better version or on-trend update. With the increased data and energy demand, we may see an increase in e-waste from AI-related data centres. Adding even more to the problem.
AI’s wider effect on the planet
This is in no way a conclusion. Nor is this post as wide as the possible positive and negative results of what AI could do. There is so much yet to happen, or that could happen.
We’re at an interesting stage here. AI is suddenly on the rise. If we’re not careful it could have hugely life-changing outcomes for human life, plant and animal life, and the planet itself. It could be either positive or negative. It could be both in different aspects of life. What we need now is governance, collaboration and a realisation that AI could be adding to the plethora of issues that damage the earth and the life that resides on it.