Extractive industries (such as oil and gas and mining) are data driven businesses and throughout the value chain, data drives decision making. Energy and Natural Resources companies have for decades been leaders in adopting and using new technologies. The mining industry’s early adoption of lidar technology from military use is one relatively recent example.
As technologies develop and ESG considerations mature, the energy transition is also increasingly data driven. Monitoring, Verification and Reporting (MVR) of greenhouse gas (GHG) emissions and reductions through IoT (the internet of things) enabled monitoring systems is a key component of understanding where we are in connection with carbon reduction targets.
AI, and particularly generative AI, are no exceptions. The natural resources sector has been an early adopter of its power in interpreting and analysing significant data sets, which are generated throughout the value chain.
As we move down the value chain, the ability of AI to support upstream, midstream and downstream activities is clear. Whether it is used as part of the optimisation of infrastructure assets, for predicative maintenance, or for modelling and operating of new assets through digital twins, AI and IoT are now at the core of any natural resources related project. Some examples throughout the value chain are:
Upstream: At the earliest stage of any natural resources project (whether oil and gas or mining) the data collected by geologists, through to the data derived from seismology, aeromagnetic surveys and drilling, needs interpretation to determine the likelihood of economically viable resources being in place. AI is also being used to help create the optimal design for wells and mines, as well as for modelling reservoir characteristics as oil and gas fields are exploited.
Midstream: In the midstream (the transportation and processing of natural resources), AI is being used for early-stage design, through to the day-to-day operation of infrastructure assets. AI combined with the use of digital twins in the design of infrastructure assets supports design and operational efficiency.
AI is also supporting companies’ understanding and prediction of demand and supply. This helps with the ability to predict and prevent equipment failures, which in turn reduces outages and helps ensure production assets remain in operation when needed. Ultimately, these data sets and AI monitoring support the ability to make investment decisions for new infrastructure, based on a clearer understanding of operations.
An example of this in operation is the balancing of power grids. With real time IoT delivered energy demand from smart meters, power transmission system operators can more readily understand power demand at any point during a day. When combined with relevant weather information, it allows the system operator to understand likely demand and have generators ready to dispatch. Integrating the intermittency of renewables (such as wind and solar) with weather information allows operators to call on other sources such as hydro, gas, coal and nuclear to meet any additional peak demands as necessary.
Downstream: In the downstream sector, the use of AI continues to be relevant. The ability to quickly analyse large, constantly changing data sets helps with health and safety in downstream facilities such as power stations, refineries and petrochemicals plants. AI also supports more efficient supply chains, and energy efficiency, by highlighting inefficiencies in the overall supply chain. By using IoT enabled equipment in new build plants, with AI systems monitoring the overall plant, operators have an increased understanding of operations. This allows fine tuning that helps to increase both operational and energy efficiency.
Trading: Trading of commodities is a key aspect of each stage of the value chain. The ability of AI to (simultaneously) analyse weather patterns, commodity price movements, historical trading activity and foreign currency fluctuations (oil and gas and metals are typically traded cross border in US dollars) provides traders with the opportunity to trade on a more informed basis. When combined with AI enabled automated trading systems, the potential benefit to traders is clear.
AI and net zero: As the different parts of the world transition to net zero at different speeds, these energy transitions can be supported by the deployment of AI. One of the simplest examples would be to monitor emissions and create data sets to show where the greatest emissions arise. Other technologies can then be deployed to mitigate these emissions.
The release of methane (the main element in natural gas) into the atmosphere is one of the largest areas of concern. Whether this is methane from flaring of associated gas, methane slip (unintended release of gas where methane is being transported or used) or leakage from poorly capped oil and gas wells, the industry can contribute significantly to the reduction of methane emissions.
Data centres: With the ever-increasing growth of data, bespoke data centres are appearing in all developed markets. The energy demands these centres command is significant; in 2022, they consumed 1–1.3% of the global final electricity demand and were responsible for 1% of GHG emissions. AI, along with other emerging technologies such as streaming, cloud gaming, blockchain and machine learning, are all poised to boost demand for data service centres. AI itself is an attractive tool to such owners looking to mitigate consumption and emissions and improve efficiencies.
What does this mean for lawyers working in the energy and natural resources sector? The convergence of technology with the physical industrial world means energy and natural resources lawyers need to understand how AI, IoT and other technologies impact the transactions they advise on.
At the simplest level, the protection and licensing of intellectual property rights (IPR) continues to be a key component of the energy and natural resources sector. Many energy companies are multi-jurisdictional owing to the location of assets and their organisational structures, so IPR are critical.
From a legal commercial point of view, the value of the IPR developed by natural resources companies is significant. All large natural resources groups have research and development teams looking to develop new technologies and deploy them operationally. AI is supporting this innovation. By learning from the data insights AI provides, natural resources companies can adapt more quickly to changing supply/demand cycles, and changes in costs, as well as the consequences of aging infrastructure, or the risk of development of new infrastructure.
In structuring any natural resources joint venture, the value that a joint venture party can bring in terms of new technologies is a key consideration. Incorporating structures to license such IP to the joint venture and capture the value of any newly created IP are key provisions of the legal documentation.
Across the entire value chain, AI’s ability to interpret large amounts of data quickly, accurately and economically, is also supporting enhanced risk management and compliance. Natural resources companies are known for working in jurisdictions that are considered high risk, with the corresponding need to remain vigilant to potential compliance issues. The ability of AI to identify patterns and anomalies – whether political, economic and/or cultural – provides compliance teams with an early warning that there may be an issue.
AI of course may have its own issues depending on how the AI system has been trained. Depending on the data sets the AI has been trained with, there may be issues with whether that data is truly representative. The known issues connected with the ability of certain generative AI products to ‘hallucinate’ is a clear risk. AI in this case should assist compliance professionals, and human decision making and oversight will be needed.
As the rapid development of AI continues, legislators are starting to catch up with the technology and are considering the need to regulate the use of AI; the EU ‘AI Act’ is expected to come into force in 2025, and in March 2023, the UK government published a white paper entitled ‘A pro-innovation approach to AI regulation’. Particularly where AI is being used by energy and natural resources companies as part of their interaction with individuals and residential consumers, legislators are likely to look to regulate the use of AI.
Of particular concern would be the use of smart meters in houses combined with the credit rating of any individual or household. Should energy companies be able to predictively manage bad debt risks by combining this type of data? And if not smart meters, then where does one draw the line with respect to data usage; indeed, could AI draw that line for us?
While the advantages of AI are real, there are a number of areas where AI, in the context of energy and natural resources, creates some distinct risks that need to be managed.
Cybersecurity: Energy security is a critical aspect of the energy trilemma. Energy infrastructure is a critical aspect of national security, and any vulnerabilities that AI might introduce could lead to legal consequences. If an energy company uses AI and subsequently suffers a data breach, it could potentially face fines or other legal actions, especially if it did not take appropriate steps to secure the data or the AI system.
Liability: If an AI system fails or causes harm, determining liability can be challenging. For example, if an AI system managing a power grid causes a blackout, it may be difficult to determine whether the fault lies with the AI's developers, the energy company using the AI, or even the users providing the data.
Transparency and explainability: The ‘black box’ problem, where the decisions made by an AI system are not clearly explainable, could lead to legal issues, especially if decisions that significantly impact customers or the public cannot be adequately justified or explained.
Regulatory compliance: Energy is a highly regulated sector, and non-compliance with existing regulations could lead to legal action. For instance, AI systems that do not comply with environmental regulations or industry-specific regulations could expose the company to legal risks.
Given these risks, having non-AI enabled redundancy built into any energy system that utilises AI would be a prudent approach. As mentioned above, the concern about some generative AI systems ‘hallucinating’ would not be acceptable in critical infrastructure. A fully autonomous AI enabled energy related system at this point in time would be unlikely to be acceptable to governments and regulators.
AI, and generative AI in particular, have their place in the energy and natural resources sector. The critical function of energy globally, however, means that humans and non-AI back up redundancy systems will continue to be needed.