Source: IEA.org. Data Accessed 30th September 2026.

The need for AI oversight is driven by several well-documented risks, one of which is bias and discrimination. AI systems trained on historical datasets may inadvertently replicate or amplify existing biases. The inequality and bias are concerning, with evidence from one of the studies suggesting that the problem is widespread. A study of 133 AI systems found that 44% showed signs of gender bias, while more than a quarter highlighted both gender and racial bias. The Organisation for Economic Co-operation and Development (OECD) AI Principles and UNESCO’s Recommendation on the Ethics of Artificial Intelligence both highlight fairness and non-discrimination as key principles for the responsible development of AI.

AI may be both the climate challenge and part of the climate solution While AI has the potential to deliver sustainability productivity gains and societal benefits, it requires energy-intensive infrastructure. Beyond electricity consumption, data centres depend on significant amounts of steel, concrete, critical minerals and advanced chips, all of which carry environmental costs. As adoption grows, the focus is shifting from whether these technologies are resource-intensive to how their resource demands can be managed sustainably without slowing innovation. Demand for financing is also increasing, with global insurance premiums linked to data centres expected to rise from US$10.6 billion today to US$24.2 billion by 2030. The central question is not how fast AI should be developed, but how society will manage the trade-off between the environmental costs and societal benefits. This mirrors long-standing debates within climate investing, where investors must weigh the short-term environmental impact of enabling technologies against their long-term contribution to decarbonisation. Renewable energy technologies, electric vehicles and battery manufacturing all require substantial resources yet are widely viewed as essential components of a lower-carbon future. AI may ultimately fall into a similar category. With the evolution of technology, some modern data centres use closed loop cooling systems which only utilise what is needed to ‘charge’ the system. Additionally, they have a carbon free power utilisation and commitment. While working with local utility companies, the power spikes are not pushed onto local consumers. Utilising such evolved technologies and business practices could potentially minimise the impact on the society while advancing the need and pace of transformation. For policymakers, investors and businesses, the challenge is ensuring that AI’s environmental footprint remains proportionate to the value it creates. This will require continued improvements in data centre efficiency, cleaner electricity generation, advances in chip design and greater transparency around energy use. If managed effectively, the environmental cost of intelligence may prove to be a worthwhile investment, particularly if AI helps accelerate scientific discovery, improve resource efficiency and support the transition to a more sustainable economy. However, achieving that outcome will depend on intentional governance, responsible deployment and a willingness to confront the trade-offs rather than ignore them.

AI may be both a force for inclusion and a driver of inequality Artificial intelligence is beginning to reshape labour markets in much the same way that previous technological revolutions transformed manufacturing and communications. However, many observers argue that AI differs from earlier innovations because it has the potential to automate cognitive tasks as well as routine physical work. The choices we make about AI now are critical. Bill Gates argues that AI may affect professions ranging from customer service and software engineering to legal support and financial analysis, creating a period of significant economic disruption unless societies prepare adequately. Historically, technological progress has created new industries and occupations even as it displaced old ones. The challenge for policymakers, educators and businesses will be ensuring that workers can adapt quickly enough to keep pace with change. Reskilling, lifelong learning and workforce mobility are likely to become increasingly important as AI systems undertake more routine analytical tasks. While AI may boost productivity and economic growth, the benefits will not be evenly distributed unless investment in skills development keeps pace with the technology. According to the World Economic Forum’s Future of Jobs research, technological change will simultaneously eliminate and create jobs, with adaptability becoming one of the most valuable workforce attributes in the years ahead. One of the most important questions surrounding AI is who ultimately benefits from its capabilities. Bill Gates has also warned that AI could become either the ‘greatest equaliser ever invented’ or the ‘worst source of injustice’, depending on how access, ownership and governance evolve. If AI-powered tools increase access to education, healthcare and financial services, they could help narrow opportunity gaps and accelerate development. Conversely, if the technology remains concentrated among a small group of countries and corporations, existing inequalities may deepen. This concern is echoed in Machines of Loving Grace, where Dario Amodei highlights the enormous benefits that advanced AI could unlock but also recognises the importance of managing the risks that accompany such concentrated technological power. The companies developing frontier AI models possess unprecedented influence over the infrastructure, data and computational resources that underpin the technology. As a result, debates around transparency, competition, accessibility and regulation are becoming increasingly central to discussions about AI’s social impact. Ensuring broad and equitable access to the benefits of AI may prove just as important as the technological breakthroughs themselves. The question, therefore, may not be whether AI is inherently good or bad, but whether it can be deployed in a manner that maximises societal value while minimising harm. The choices society makes today will determine whether AI becomes a force that broadens opportunity or one that exacerbates inequality. Like many transformative technologies before it, AI presents a complex trade-off: its long-term value will depend less on the technology itself and more on how responsibly it is governed, distributed and applied.

Figure 2: Total job growth and loss