Smart AI policies that drive energy innovation while tackling emissions risks 

Since the launch of the GPT series in 2022 by OpenAI, AI has led to new possibilities beyond detecting patterns in data and using historical knowledge to predict future outcomes. By now, global technology companies and energy communities are not only aware of the potential of artificial intelligence to optimise energy grids and improve efficiency across industries, but they are also heavily investing in AI-driven solutions to enhance operations, reduce emissions, and drive innovation.

For example, BP recently partnered with Palantir to leverage AI to optimise its oil and gas operations, while Shell is actively integrating AI to transform its energy sector. Additionally, Saudi Aramco’s venture arm has allocated $100 million for AI investments, underscoring the industry's strategic shift toward AI-powered energy solutions. While these investments signal a promising future for AI in energy, the systems responsible for training and operating AI models are among the most energy-intensive technologies ever developed. 

This breakthrough in AI technologies is expected to drive up carbon emissions as demand surges, with more tech companies entering a space once dominated by OpenAI’s ChatGPT, Google’s Bard, and Microsoft’s Co-Pilot. However, the latest entry of China’s Deepseek, an open-source AI model, aims to democratise access and foster technological inclusion by reducing entry barriers. 

Current risks associated with AI dominance

Although AI’s use cases are well-intended, they sometimes fail unintentionally, raising concerns often related to content, safety, resources, ethics, and accuracy risks, to mention a few. For example, a content risk can occur when weather data from a specific region is used to train an AI. The data from that region could skew energy forecasts or misallocate resources like solar or wind capacity, leading to inefficient power systems. Safety and security risks emerge, such as cyber-attacks targeting AI systems that are being integrated into the power sector, alongside resource risks like strained electrical grids or heightened greenhouse gas emissions from energy-intensive models. 

These challenges leave our society grappling with questions like: 

• How do we foster transparency in AI development? 

• Should efficiency take precedence over sustainability during their creation, or can we balance both?

• How do we repurpose the AI hardware? 

• How can we mitigate the pollution from the use of this AI hardware?How might tech companies, energy experts, and governments collaborate on policies that drive innovation while safeguarding societal well-being?

Existing policies for AI dominance and energy security are fragmented

 These concerns have yet to be fully addressed. As with many technological breakthroughs, leading companies prioritise first-mover advantage to establish dominance and attract significant investment - often regardless of whether policies are in place to guide the ethical and responsible use of these technologies.

Within the energy sector, the same concern exists about the lack of clear policies that can shape the development of AI for optimising energy efficiency in different jurisdictions like the EU, China and the U.S., leaving businesses uncertain about the future of investments and development while governments struggle to regulate emerging AI models. 

Since 2020, the EU, US, and China have been developing AI regulatory frameworks, though none specifically address AI and energy. China, through the cyberspace administration of China (CAC), has taken the lead on introducing new legislations to regulate AI technologies, followed by the European Union’s Artificial Intelligence (AI) Act which entered into force on 1 August 2024. In Europe’s AI Act, there seem to be requirements for systems to be designed with the capability of logging their energy consumption; while researchers are designing specialised hardware like accelerators and 3D chips for improved performance. 

However, in the U.S., there is no national policy framework for regulating the power to train and operate an AI model. Under the Biden administration, executive orders emphasised responsible AI development, including measures for transparency, accountability, and risk mitigation, with some attention to data center energy needs, whereas current policies focus on deregulation to promote AI innovation without energy-specific mandates. 

This absence of standardised policy frameworks for AI-driven energy resilience leaves room for potential U.S. regulation. Therefore, establishing clear regulations could ensure consistency in energy usage monitoring and promote best practices, reducing negative environmental impacts, perhaps by mirroring Europe’s logging requirements or incentivising efficient hardware. However, any shift would likely prioritise competitiveness over strict sustainability mandates under the current administration.

Given the ongoing discussion on what the policy approach should be concerning these current challenges and market dynamics, a relevant concern is how AI delivers on its promise as a tool for optimising energy grids and improving industrial efficiency while addressing its environmental footprint. How should institutions weigh the business benefits of AI against the environmental costs? Should market forces drive this balance, or is there a role for regulation to ensure carbon-intensive models are retired in favour of sustainable alternatives? How can Industry leaders, policymakers, and researchers collaborate to ensure AI innovation aligns with energy security goals while delivering measurable economic and environmental value?

A need for pragmatic policies to address current concerns

To realise the full potential of AI in optimising energy development, there is a need for cross-sectoral collaboration between business leaders, governments, and researchers; however, regulating AI for energy should be done with caution because, as with most technological developments, one recurring question is whether regulation should anticipate or follow technological advancement. In the past, policy guidelines and regulatory frameworks have tended to follow innovation, responding only after technology disrupts society. Both China's and the EU’s legislation aims to preempt risks and ensure that AI development aligns with society's needs. This anticipatory model seeks to protect individuals from potential risks, such as data breaches, cybersecurity breaches, and algorithmic bias as earlier mentioned in this article, while fostering innovation.

Equally, special consideration should be given to energy and the environment. The current AI landscape has witnessed massive investment in AI startups, experimental business models, and increasingly speculative applications. Given this, overregulation may hinder technological progress, among other things. On the other hand, under-regulation may expose society to significant harm, as seen in cases like the rise of carbon emissions or instances where researchers who are designing specialised hardware for computing are unknowingly creating more paths for energy consumption. 

Industry momentum

Case studies underscore AI’s transformative potential in the development of energy. Some reports suggest that energy companies are keen on adopting AI to optimise their operations but are wary of overregulation stifling innovation; tech firms advocate for policies that encourage the development of energy-efficient AI technologies while providing a balanced regulatory environment, regulators focus on creating frameworks that ensure AI's safe and ethical use, addressing concerns like data privacy and cybersecurity. Meanwhile, environmental groups emphasise the need for stringent regulations to minimise AI's carbon footprint and protect natural resources. All these concerns require careful consideration of implementing policies that balance AI innovation and energy development.

Policy considerations

What type of policies can government and business leaders establish to address society concerns, advance AI’s development for energy without exacerbating the environmental crisis, and attract investments as AI development continues? Getting the right policies will be key in shaping how AI integrates into the energy sector and society.

Engaging business leaders

Very few governments have policies that address AI’s role in energy. Although China and the EU have a few regulations that govern the development of AI as a whole, the starting point would be to assess these policies and understand different aspects of energy and environment that could be considered as these models are being developed. To start, energy and technology leaders, and other stakeholders must align on what business goal AI in energy would help to achieve and business leaders should actively engage with policymakers by partnering with governments to create public-private initiatives for AI-powered energy security. Equally, business leaders should advocate for regulatory sandboxes that allow real-world testing of AI solutions in energy grids – this would help minimise potential risks associated with its development, ensure transparency, and support trustworthy AI ecosystems. Engaging in the policy conversation not only ensures a favorable regulatory landscape but also solidifies a company’s role as an industry pioneer.

Government’s leadership

On the other hand, governments play a crucial role in creating an ecosystem where AI-driven energy solutions can thrive. The government may start by mandating technology companies to create an energy and environment office to monitor the consumption or energy use of these technologies and provide feedback to help address it; government and philanthropy organisations should fund research development of AI to lead to energy security; training specialists that could provide standards developing AI technologies that puts sustainability as part of the AI development process and not as an afterthought, providing people with training and skills needed to understand the development of AI models data experts or data protection officer; also, government must mandate the disclosure of the data of energy consumed while AI technologies are being developed.

Conclusion

There is a clear use case for AI to catalyse energy development; however, this will be realised only if the current risks are addressed. Current concerns and risks must be addressed, including the need for responsible and ethical development of AI technologies. 

As the next-generation algorithm and chips make their way into the market, businesses cannot wait for the government to develop policies; the government needs the support of experts in establishing policy guidelines that will aid the development of AI to optimise energy development while prioritising sustainability. These policies include but are not limited to: 

1. Establishing clear benchmarks for the energy efficiency of AI systems and requiring compliance with these standards to ensure that AI technologies developed and deployed are optimised for low energy consumption; 

2. Setting ethical AI guidelines for the deployment of AI technologies in the energy sector. These guidelines should address issues such as data privacy, algorithmic bias, and the environmental impact of AI systems; 

3. Education and training: this would equip the workforce with the skills needed to develop and implement energy-efficient AI technologies and 

4. Conducting environmental impact assessments: this ensure that companies developing AI identify potential environmental risks and mitigate them before deployment

Success will require collaborative efforts among government, business leaders, and institutions to ensure that AI's revolutionary potential serves the broader goals of energy equity, social progress, and economic development.

illuminem Voices is a democratic space presenting the thoughts and opinions of leading Sustainability & Energy writers, their opinions do not necessarily represent those of illuminem.