How energy-efficient is AI? Torn between cybersecurity benefits and increasing power demand

Artificial Intelligence is getting traction faster than we’d like ― considering the lack of legal framework around the subject. The discrepancy between the fast-evolving technology and the necessity for ethical regulation makes AI challenging to control. 

That’s why AI can be controversial from a few perspectives, such as eco-friendly ones. While worldwide companies already leverage these systems and benefit from automated responses in cybersecurity, and using them can strengthen vulnerability management, these tools demand more and more energy. 

The carbon and water footprint used for training and running massive amounts of data is increasing. At the same time, the daily queries these AI models receive also require a lot of electricity, while training them needs water― an important element in preventing servers from overheating. 

So, how can the world find the balance between the two Ai-related issues?

AI for protecting companies from cybersecurity threats 

Although modern technology introduces stronger safety measures, companies are still not fully prepared to withstand risks. According to the latest statistics provided by the UK’s Department for Science, Innovation, & Technology, over four in ten businesses and three in ten charities experienced different types of cybersecurity breaches in the past year. While there has been considerable improvement in how small firms practice cyber hygiene, fewer companies review the risks coming from suppliers, for example. 

Unfortunately, consumers are also in line for exposure, besides companies. They might be exposed to different types of breaches, whether concerning GDPR (General Data Protection Regulation), email addresses, or bank data. Advisors from DataBreachCompensationExpert address how processors that fail to protect personal information can be sued for compensation, as 

they breached their accountability to adhere to the seven core principles of personal data. 

However, the possibility of data breaches and cyber threats can be lowered with the help of AI. Its modern capabilities bring speed, scalability, and precision in the cyberspace by:

• Detecting anomalies and flagging deviations from the norms; 
• Analysing vast datasets to detect patterns; 
• Using machine learning to forecast threats; 
• Automating responses to cyberattacks; 

But what’s the cost for all these operations? 

AI heavily relies on algorithms and data, which require continuous training and improvement. For that, the data centres in which AI happens require significant coal and oil resources for computational power to drive AI algorithms to work. Rare minerals like mercury and zinc are needed to deliver the machines that train and run AI models. 

There are many other environmental concerns regarding AI. The continuous growth of fossil fuel use is linked with the rise of data centres throughout the globe, especially in the US. One ChatGPT search requires the energy equivalent of ten Google Searches, while Microsoft’s data centres might’ve needed about 700,000 litres of water to cool the machines used for training AI models. 

But there’s more to it. Environmental groups believe AI only contributes to climate misinformation since the tool has the potential to help us fight climate change, but these operations consume an unprecedented amount of energy. 

The hype surrounding AI has made it difficult to control the energy requirements for running it. Alternative resources for AI models, including solar and wind, should replace the demand for natural gas, but this also requires an expanded and strengthened infrastructure. 

How can we make technology more sustainable?

Technology has indeed boosted the world’s economies, leading us to a better world, but we’ve got ahead of ourselves when it comes to preserving natural resources. Earth’s resources are finite, but we have used them like they’re unlimited due to the lack of awareness around the subject. 

But that doesn’t mean returning to the past. Being more sustainable and tech-savvy includes:

• Choosing energy-efficient gadgets that consume less power; 
• Going paperless by using more digital documents; 
• Learning to repair instead of replace; 
• Supporting sustainable tech initiatives; 

However, when it comes to the industries that have made a significant contribution to climate change, things are a bit different. Among some of the many changes they can approach, companies should:

• Reduce their dependencies on natural resources and switch to renewable energies; 
• Control their consumption and monitor consumption in power plants; 
• Minimise their production of waste and manage it better; 
• Optimise industrial production to anticipate incidents and improve equipment effectiveness; 

Can AI truly contribute to a better future? 

The use of AI is like a double-edged sword for now. If the economic benefits outweigh the environmental challenges AI poses, the technology could mitigate emissions reductions since the World Economic Forum estimates that GHG emissions will remain below 1.5% in the following ten years. 

Experts are now working on specialised hardware that could deliver much more performance than current AI technologies at the cost of 25 times less energy. Accelerators, 3D chips, and new chip cooling techniques will make AI less detrimental to nature. This applies to data centres as well, as specialists explore new and more sustainable ways to cool them down. 

AI’s potential in powering up sustainable development also includes:

• Organisations like Stream Ocean leverage AI and machine learning to monitor marine diversity and use advanced data analytics to address SDG 14 (Conserve and sustainably use the oceans, seas and marine resources for sustainable development);
• Investors can assess climate risk exposure through companies like BlackRock, tracking the carbon footprint of investments; 
• Impact Management & Measurement (IMM). We need accountability and transparency in measuring progress, which is what the European Commission (EC) does to manage the progress of SDG 6 (Ensure availability and sustainable management of water and sanitation for all); 

Final considerations 

AI is one of the most talked-about subjects at the moment, but so is sustainability and climate change. These two industries need to work together to withstand the effects of ever-growing technology, which is why AI can help, but not in its current state. Therefore, organisations must use it responsibly and with improvement in mind so that they can support all Sustainable Development Goals for a fairer, healthier, and modern world. Using AI through renewable energy is the solution.