Autonomous Electricity Market Agents

1. Introduction

Autonomous Electricity Market Agents are software-based systems that can independently perform activities in electricity markets using programmed rules, artificial intelligence (AI), machine learning and real-time data. They may act for generators, traders, aggregators, storage operators, consumers or other market participants.

An autonomous agent may monitor electricity prices, forecast demand, submit bids, purchase or sell electricity, manage battery storage and adjust transactions automatically. This technology can make electricity markets faster and more responsive, especially as renewable energy and distributed generation increase.

However, an autonomous software agent does not automatically become a legal person. Its actions generally remain attributable to the company, trader, generator or other legal entity that operates or authorises it. Therefore, energy law must determine who is responsible when an autonomous agent makes an unlawful or harmful decision.

2. Meaning and Functions

An autonomous electricity-market agent can perform several functions:

Price monitoring: continuously observes electricity-market prices.

Demand forecasting: predicts future electricity consumption.

Bid submission: automatically submits generation or purchasing bids.

Energy trading: buys or sells electricity according to programmed conditions.

Storage management: decides when batteries should charge or discharge.

Risk management: identifies abnormal prices or trading conditions.

Compliance monitoring: checks whether transactions comply with market rules.

For example, an energy-storage operator could use an autonomous agent to charge a battery when electricity prices are low and discharge it when prices increase. The agent could make thousands of decisions without direct human instructions.

3. Legal Framework in South Africa

The Electricity Regulation Act 4 of 2006 (ERA) provides the central legal framework for electricity generation, transmission, distribution, trading and related activities. NERSA has important responsibilities concerning licensing, regulation, compliance and electricity-market development.

NERSA has also been developing Electricity Trading Rules to provide a clearer framework for electricity trading as the South African electricity market evolves. (nersa.org.za)

Autonomous agents must therefore operate within applicable licence conditions, market rules, grid codes, contractual obligations and competition law.

An important legal principle is that automation cannot create independent regulatory authority. A company cannot escape liability simply because an algorithm made the transaction.

4. Relevant Case Laws

Eskom Holdings v Vaal River Development Association (2022)

In Eskom Holdings SOC Ltd v Vaal River Development Association, the Constitutional Court considered the statutory and constitutional framework governing electricity supply, including Eskom's responsibilities and NERSA's regulatory role. (www3.saflii.org)

The case is relevant to autonomous market agents because electricity-market activity takes place within a regulated public framework. Automated trading cannot be separated from the legal duties imposed on electricity-sector participants.

Eskom Holdings v Lekwa Ratepayers Association (2022)

The Supreme Court of Appeal considered the relationship between Eskom's bulk electricity supply and municipal distribution responsibilities in Eskom Holdings v Lekwa Ratepayers Association. (saflii.org)

The case demonstrates that electricity-market roles are legally structured. An autonomous agent operating for a municipal distributor, generator or trader must remain within the authority and responsibilities of its principal.

Eskom Holdings v Sonae Arauco (2024)

In Eskom Holdings v Sonae Arauco, the Supreme Court of Appeal considered the South African Grid Code and the System Operator's responsibilities concerning electricity-system reliability and remedial action. (saflii.org)

This is relevant because autonomous market agents must not focus only on financial profit. Their actions can affect system stability. Trading and dispatch decisions must therefore remain consistent with applicable grid requirements.

Pharmaceutical Manufacturers Association v President (2000)

The Constitutional Court in Pharmaceutical Manufacturers Association v President established that public power must be exercised lawfully and rationally.

The principle is relevant where autonomous agents are used by public bodies or regulators. An algorithm cannot exercise a public power beyond the authority granted by law.

5. Competition and Market-Manipulation Issues

Autonomous agents create important competition-law risks.

If many agents use similar algorithms, they could potentially produce coordinated market behaviour even without direct communication between human traders. Agents might also respond rapidly to competitors' prices and create unstable pricing patterns.

Other risks include:

automated market manipulation;

false or misleading bids;

excessive trading;

strategic withdrawal of generation;

discriminatory access;

algorithmic collusion; and

manipulation of congestion or balancing markets.

Therefore, electricity-market rules should require appropriate monitoring and auditability.

6. Accountability and Liability

A central question is: Who is responsible for an autonomous agent?

Possible responsible parties include:

the electricity trader;

the generator;

the aggregator;

the software owner;

the system operator; or

another legally responsible entity.

The safest legal approach is to maintain a clear human or corporate chain of responsibility.

Market participants should maintain records showing the agent's programming, data sources, decisions, transactions and human supervision.

7. Consumer and Data Protection

Autonomous agents may process detailed consumer information, including electricity consumption and transaction histories. Data-processing obligations therefore become important.

Consumers should receive understandable information about automated decisions that significantly affect their electricity services or financial obligations.

Cybersecurity is equally important. If an attacker takes control of a market agent, the attacker could potentially manipulate electricity purchases, sales or storage decisions.

8. Conclusion

Autonomous Electricity Market Agents can transform electricity trading by allowing market participants to react automatically to prices, demand, renewable generation, grid conditions and storage opportunities.

Their development must nevertheless remain within the legal framework created by the Electricity Regulation Act, NERSA rules, Grid Code, competition law, contractual principles and administrative law.

The cases of Eskom v Vaal River Development Association, Eskom v Lekwa Ratepayers Association, Eskom v Sonae Arauco and Pharmaceutical Manufacturers demonstrate the importance of statutory authority, system reliability and accountability.

The key principle should be “autonomous operation without autonomous legal responsibility.” An algorithm may act independently in technical terms, but the licensed entity or institution responsible for it must remain accountable for compliance, market conduct, cybersecurity, consumer protection and the consequences of its actions.