AI Has a Real Environmental Impact—Here’s How Designers Are Handling It

AI Has a Real Environmental Impact—Here’s How Designers Are Handling It

In recent years, the use of artificial intelligence (AI) has gained significant momentum in various industries. From personalized recommendations to news apps, AI has become an integral part of our daily lives. However, as AI continues to evolve, there is a growing concern about its environmental impact.

Jarno Koponen, head of AI and personalization at Yle, Finland's national broadcasting company, has been researching the use of AI in the digital media landscape. He points out that while consumer applications of AI often use specific and narrow AI models, they still require significant computer processing power. This raises questions about the emissions generated by AI and the need to address this issue in the broader consciousness.

Koponen believes that sustainable design and development principles need to be updated to account for the era of AI applications. Currently, the predominant metric of success in the ML research community is accuracy. However, striving for higher accuracy often comes at the cost of increased computing power and energy requirements. Koponen suggests that finding a balance between cost and accuracy is crucial.

This idea is echoed by Schumann, who emphasizes the importance of cost-efficiency in AI development. The difference between 80% and 90% accuracy can have a significant impact on energy consumption, with the latter requiring much more computing power. Striving for perfect accuracy may not always be necessary, especially in applications where the stakes are not as high. Instead, AI should be used to assist, rather than replace, humans in the equation.

Otsuka, a type designer, shares a similar perspective. In his forthcoming project, he plans to use machine learning (ML) to assist type designers in kerning new typefaces. Instead of automating the kerning process, the tool will provide suggestions based on what it has learned from reading other typefaces. The aim is not to achieve perfection but to speed up the laborious process for designers. Otsuka believes that running the tool long enough will provide designers with a starting point.

While imprecision may not be suitable for every AI application, it highlights the need to consider the human factor in AI design. AI should be a tool that aids humans, rather than replacing them entirely. This perspective aligns with the idea that AI should be used in a way that benefits human livelihoods.

However, designers are currently limited by the capabilities of consumer hardware. In the creation of the visual identity for an exhibition on artificial intelligence, Process Studio trained an AI model to generate its own emojis. The model was fed a dataset of commonly used emojis, and the result was a unique set of emojis created by the AI. This example showcases the creative potential of AI in design.

In conclusion, AI's environmental impact is a growing concern that designers are starting to address. The push for accuracy in AI development has led to increased energy consumption and emissions. To mitigate these issues, designers are exploring the concept of imprecision and the use of AI as a tool to assist, rather than replace, humans. Additionally, limitations in consumer hardware are pushing designers to think creatively and find new ways to leverage AI's potential.

Three actionable advice for designers in handling the environmental impact of AI:

• 1. Consider the trade-off between accuracy and energy consumption. Strive for a balance that achieves the desired outcome while minimizing environmental impact.

• 2. Embrace imprecision in AI design where appropriate. Not every application requires perfect accuracy, and allowing for some level of imprecision can reduce energy requirements.

• 3. Use AI as a tool to aid human creativity and productivity, rather than replacing human involvement entirely. AI should be seen as a partner rather than a substitute.

By incorporating these principles into AI design, we can pave the way for a more sustainable and environmentally friendly future.