Green AI: Deloitte Outlines The Need to Make Tech Green

As AI fervour reaches fever pitch, with enterprises adopting on mass and individuals using the tech independently at home and work, the problem of how this is all being powered is an issue that needs to be addressed.

A recent report by Deloitte examines the ecological ramifications of AI's swift growth. One of the critical issues highlighted is the projected tripling of global electricity consumption used by data centres to power this over the next decade.

Consequently, Deloitte emphasises the necessity for AI development to align with global climate goals to mitigate its environmental impact.

The carbon cost of AI

Approximately 0.3% of global greenhouse gas emissions are attributed to data centres, which serve as a crucial infrastructure for the advancement of AI, a figure due to increase as more move from storage to AI computations.

Moreover, the energy-intensive processors such as GPUs and TPUs, which consume more power than conventional computing devices, are essential for training and operating AI models.

This reliance further exacerbates the environmental impact associated with AI technologies.

The swift adoption of AI across various sectors poses significant risks, particularly in areas still dependent on fossil fuels for electricity.

This trend leads to increased computational demands and higher electricity usage.

Without the establishment of regulatory frameworks and the promotion of collaborative efforts across the industry, there is a genuine concern that AI could hinder our progress towards achieving sustainable development goals.

Powering AI 

Deloitte Global emphasises the critical need for stakeholders to work together in developing sustainable solutions that align AI advancement with global sustainability objectives. 

The research focused on three fundamental questions regarding the environmental effects of AI and its energy consumption.

Data centres currently consume 380 TWh of electricity globally, accounting for approximately 1.4% of total global electricity usage.

Projections indicate that this figure could rise to 1,000 TWh by 2030 and reach 2,000 TWh by 2050, representing around 3% of global energy consumption.

If the demand for AI continues to grow at its current rate without any improvements in efficiency, Deloitte warns that energy consumption could surpass 3,500 TWh by 2050.

To address the detrimental environmental impacts associated with the expansion of AI, the report underscores the necessity of implementing key strategies for sustainable AI development.

"The potential of AI to reduce waste and optimise supply chains is great but so is its energy consumption," says Geoff Tuff, Principal of Deloitte.

"I haven't yet seen, at scale, the rational conversation about balancing AI model sophistication based on dominant use with energy intensity and trade-offs."

Getting to 'green' AI

Deloitte Global's report, "Powering AI," underscores the necessity of adopting 'green AI.' This concept aims to balance rapid technological progress with environmental accountability.

By developing AI technologies that optimise resource usage, align with sustainability objectives, and enhance energy efficiency, it is possible to ensure that future AI advancements support our sustainability targets and positively impact the climate.

Green AI will improve hardware and data centre operations to maximise energy efficiency, enhance transparency through standardised energy usage reporting, and prioritise renewable energy sources for powering AI initiatives.

Central to this initiative are green data centres, which operate with lower energy consumption and carbon footprints by using advanced technologies and sustainable materials.

Heat recovery is another important aspect, capturing waste heat from computing systems for reuse, thus improving energy efficiency.

Efficiency cooling systems also play a crucial role by maintaining optimal temperatures with reduced energy use compared to traditional methods.

Sustainable computing practices focus on the entire lifecycle of computing resources, promoting eco-friendly materials and responsible recycling of outdated equipment, and resource optimisation further enhances efficiency by maximising resource use and streamlining processes.

Overall, a strong emphasis on energy efficiency across these practices ensures that green AI contributes positively to sustainability goals, fostering a future where technological advancement aligns with environmental responsibility.

Ultimately, the emissions AI creates means that the future of AI must be approached with a commitment to sustainability. 

Explore the latest edition of AI Magazine and be part of the conversation at our global conference series, Tech & AI LIVE. 

Discover all our upcoming events and secure your tickets today. 

AI Magazine is a BizClik brand