Consumer Energy Resources: AI and automation

Tell me in two minutes

Consumer energy resources (CERs) are consumer-owned devices that generate, store, or manage electricity. These devices include smart meters, rooftop solar, batteries, electric vehicles and smart inverters. The integration of Artificial Intelligence (AI) in CERs has the potential to revolutionise the energy sector by enhancing traditional methods of CER operation and management. By leveraging optimisation algorithms, historical and real-time monitoring and advanced analytics, AI-enabled CERs could offer immense value for all CER users by optimising energy consumption and ultimately reducing energy costs.

In this article, we discuss the potential for AI to be deployed in CERs, risks to watch out for, and an update on the regulation of AI in Australia.

This is the third article in our five-part series on CER, where we are exploring the emerging opportunities and challenges associated with the uptake of CER in Australia from a tech law perspective.

You can read our first article (Overview of CER) here, and our second article (Data and Privacy) here.

How can AI be applied in CER systems?

Machine Learning

Machine learning is a subset of AI that uses algorithms and statistical models to enable a system to use data to learn and improve their performance on tasks, without being explicitly programmed to perform those tasks. You can read more on machine learning here.

Machine learning can be used to analyse historical and real-time data from CERs to identify patterns, detect anomalies, and make predictions. This can be used to optimise the operation of CERs and improve their performance, including forecasting energy consumption and generation output (especially intermittent output from renewable energy sources), demand response optimisation, and fault prediction.

Reinforcement Learning

Reinforcement Learning (RL) is a subset of AI that trains systems to make decisions that achieve optimal results. You can read more on RL here.

RL algorithms can learn optimal control policies for CERs by interacting with and receiving feedback from the systems environment. RL can be used to prioritise various objectives like minimising the energy costs for consumers and grid stability.

Data Analytics

Data analytics refers to a range of tools and techniques that are used to derive meaningful insights from data.  

Data analytics techniques, such as data clustering, time series analysis, and anomaly detection, can be applied to CER data (such as generation and consumption, power quality, grid performance, telemetry and pricing) to understand trends and drive informed decision-making by consumers and CER system operators. You can read more on data and CERs here.

Case studies that combine AI and CER

Benefits of AI application in CER systems

AI-enabled CERs, and using AI in conjunction with CERs and CER data, can facilitate efficiencies in power generation, distribution and use, as well as efficiently integrating traditional and renewable energy sources into the grid. At scale and effectively coordinated, it could accelerate the transition towards a cost effective and renewable energy landscape.

At a network level, AI systems can conduct historical and real-time analysis of CER generation outputs, local grid conditions and demand patterns. AI systems can also conduct, or assist human decision makers in, load forecasting and adjusting energy dispatch or signalling for consumers to adjust their consumption. By optimising the balance between supply and demand, AI systems can help to prevent grid overload and blackouts during peak demand periods. AI systems could also mitigate the instability of renewable weather-dependent generation CERs (like solar and wind) by integrating traditional energy sources or firmed renewable sources when necessary. From a maintenance perspective, AI-enabled prediction and detection of abnormal events can facilitate predictive fault finding to reduce supply interruptions and avoid the need for urgent, costly infrastructure upgrades.

At the individual level, CER devices such as smart meters and other digitally linked household devices can enable consumers to both better understand and adjust their own energy usage to maximise energy savings.

Commercially available devices with AI functionality can also allow consumers to automate decision-making. For example, turning off non-essential devices to reduce energy consumption, or scheduling device activation (like EV charging) to align with hours of lowered grid demand. These behaviours have the potential for cost savings, depending on the incentives offered by local energy retailers.

Challenges and considerations in AI integration for CER systems

Integrating AI into CER systems has the potential to introduce certain risks for consideration. You can read more about data and privacy in CER systems here.

Legal and Regulatory Implications

CER participants (particularly CER manufacturers and suppliers) should be cognisant that AI (including in relation to CER) is not ‘unregulated’ in Australia. Competition and consumer law, privacy law, copyright law and consumer protection provisions contained in the National Energy Customer Framework and other legislation already govern AI. Liability can, to a certain extent, be managed privately through contracts between CER participants, but CER participants should ensure that they are mindful of the existing patchwork of regulation that applies to AI, if they are planning to use this technology.

Nevertheless, in our review of the Government’s interim response to the 2023 Safe and Responsible AI in Australia consultation (which can be found here), we discussed that the Government acknowledged in its interim response that existing laws do not adequately address the risks presented by AI. The Government indicated that it would consider the appropriateness of mandatory safeguards for AI development and use, and look to leverage mandatory guardrails in any high-risk setting where those guardrails already exist.

The Government will also seek to address AI risks through a suite of legal reforms, including by strengthening existing privacy laws. In its 2022 review of the Privacy Act, the Government proposed that ‘privacy policies should set out the types of personal information that will be used in substantially automated decisions that have a legal or similarly significant effect on an individual’s rights’. After undergoing significant public consultation, draft reforms to the Privacy Act are expected in September 2024.

The importance of having an AI governance framework

In its interim response, the Government noted that its approach to regulation would be guided by certain principles: risk-based, balanced and proportionate, being a trusted international partner in addressing AI risks, collaborative and transparent with the public, and prioritising people and communities. The Government will also further consult on introducing regulation, focusing on testing, transparency and accountability to prevent harms. CER participants should consider if they are developing and deploying their products in a safe and responsible way, and how their activities measure against the Government’s regulatory principles. CER participants should:

• establish a risk-based AI governance framework that defines acceptable risk levels appropriate to each organisation’s activities
• map their activities against their AI governance framework, and
• keep their AI governance framework up to date with current industry codes or standards (eg the AI Safety Standard, to be released following development with industry).

We will be publishing a new article on AI governance soon, so keep an eye out.     

The next article in this series:

In the next article in this series, we will be discussing the intersection between cyber (including cyber risks and cyber preparedness) and Consumer Energy Resources.

In the meantime, you can read our other AI alerts here, and view our map of AI regulation in Australia here.