This article was originally published in Law360 on December 12, 2023, and is republished here with permission.
President Joe Biden’s recent executive order to manage risks associated with artificial intelligence is intended to establish new standards governing AI safety, security and privacy concerns, while promoting innovation and competition.
Issued on Oct. 30, the modest — for federal regulatory standards — 53-page executive order on the safe, secure and trustworthy development and use of artificial intelligence will have the most immediate effect on organizations and business directly developing AI and AI tools.
But, the executive order previews a breathtaking scope for AI to potentially enter our daily life, spanning advanced technology and research, housing, health care,[1] education, the labor force,[2] climate change, immigration, infrastructure and government procurement.
Naturally, the executive order could have reverberating implications for many aspects of the U.S. economy, including the energy sector.
Executive Order Guiding Principles and Regulatory Mandate
The executive order frames the issue of AI and the Biden administration’s response as being led by eight guiding principles and priorities that can be characterized as follows:
1. Ensure safety and security in AI development and deployment;
2. Position the U.S. for economic leadership in AI problem-solving;
3. Support a domestic American workforce;
4. Advance equity and civil rights;
5. Uphold and expand consumer protections;
6. Protect privacy and civil liberties;
7. Develop the federal government’s capacity to use AI; and
8. Coordinate the international community’s response to AI.
The landmark executive order provides a road map for energy sector companies to begin addressing larger general public concerns regarding the growth and use of AI.
The executive order specifically encourages independent regulatory agencies to consider using their full range of authorities, including rulemaking and revising existing regulations and guidance that apply to AI, to protect American consumers from fraud, discrimination, financial instability, threats to privacy and other risks that may arise from the use of AI.
AI in an Evolving Energy Grid
The executive order calls for efforts to use AI to improve the planning, permitting, investment and operations for electrical grid infrastructure to enable the provision of clean electric power.
In the energy sector alone, AI is already being deployed in applications to interact with potential residential solar and energy storage customers to process consumer data and accelerate screening for consumer energy products. This offers companies the prospect of greater efficiency in otherwise costly customer acquisition and site-by-site system design.
Grid operators, utilities and energy market participants are increasingly looking to AI-based tools to more efficiently manage, automate and transact in a distributed, interconnected and instantaneous energy grid. This includes merchant energy markets that can be subject to volatility during extreme weather events and natural disasters.
This reveals a complexity feedback loop currently underway whereby the evolving ecosystem enables the introduction of new variables, which then puts additional technological and computational pressures on the existing system to keep up, thereby driving even great reliance on advanced machine learning.
For example, the historical power grid has stretched and accommodated significant quantities of intermittent and distributed renewables, requiring new technologies like energy storage and demand response to smooth out volatility in supply and demand. But this deployment then places greater demands on grid operators to optimize their models and more accurately forecast grid needs.
The challenge is further compounded by adjacent market forces like electric vehicles that will be a significant load factor and increasingly interact bidirectionally with the grid.
Initial Energy Sector Impact and Counsel Considerations
Within the energy sector, the executive order instructs the U.S. Department of Energy to address AI systems’ threats to critical infrastructure by coordinating with the National Institute of Standards and Technology to set rigorous standards, subject to so-called red team testing that simulates worst-case scenarios to ensure safe deployment of AI systems within energy infrastructure, all prior to public release of the technology.
The Department of Energy is now beginning to establish rules and regulatory compliance in accordance with this executive order, a process that will take at least the next nine months.
In the meantime, in-house counsel may wish to work with business lines and executive teams to consider implementing their own AI governance process to prepare for forthcoming legal requirements, especially if their subsector is specifically identified in the executive order.
Given the breadth of issues, consider identifying a primary point of contact in the company on AI issues, and take an inventory of company tools, processes and practices that could be affected. This may include AI tools provided by third-party vendors.
Once AI elements are identified, departments using AI may want to consider taking steps to ensure that their technology is auditable or transparent, preserving the ability to explain how their AI solutions operate and what data is required.
In an effort to get ahead of the forthcoming regulations, companies would likely want to focus at the outset on maintaining the privacy and security of any data used by the AI technology, for example, by complying with the above-referenced NIST data security standards.
In addition, energy sector companies that utilize AI should also be prepared to demonstrate how their AI models are trained while preserving the privacy of the training data. Companies should consider what transparency obligations are included or omitted in any third-party vendor contracts, especially considering that AI developers will be fiercely protective of their proprietary information in the current competitive AI field.
Energy sector companies that develop AI technology that may be a serious risk to national security should also be prepared to share their safety test results and any other critical information with the U.S. government. Any such companies should be prepared to explain how their AI includes or is surrounded by adequate protections against AI-enabled fraud and deception.
Further, those developing AI for energy sector companies would want to be able to demonstrate authentication features that would make it clear that any AI-generated communications are authentic.
Lastly, growing companies should be prepared to answer due diligence questions in capital-raising and mergers and acquisitions contexts about steps they have taken to preserve and protect their AI intellectual property, and also communicate their exposure to potential future regulatory actions.
To the extent federal agencies do pursue new regulations, investors and financing parties will want to see adequate measures put in place to ensure compliance.
An Uncertain Future Moving at Lightning Speed
As these systems become more complex and system decision making becomes further removed from direct human intervention, innovators, sponsors and financing parties will need to further evaluate the allocation of liability and risk for artificial intelligence.
Losses or other negative outcomes from an AI tool could be viewed by the user as reflecting system defects, whereas the developer may lay blame at the feet of the user for poor AI prompting or other more fundamental deficiencies in the user’s ability to wield a powerful new tool.
But as the private sector only begins to wrestle with these novel legal issues, the Biden administration’s executive order could prompt regulators to pursue new rules or restrictions for any identified emerging risks to consumers’ individual privacy, energy market vulnerabilities, and community infrastructure weaknesses.
Thoughtful execution of the executive order will be critical to achieve the intended balance of security and privacy interests with the opportunities arising from the safe and effective use of AI to advance economic growth in the energy sector.
