Autonomous Dispute Resolution Systems in Energy Markets

1. Introduction

Autonomous Dispute Resolution Systems (ADRS) are technology-based systems that use artificial intelligence (AI), algorithms, smart contracts, automated evidence analysis and digital communication to assist in resolving disputes in energy markets.

Energy disputes may arise between generators, traders, distributors, municipalities, electricity consumers, system operators and regulators. Common disputes concern electricity tariffs, billing, supply interruptions, contracts, grid access, market transactions, payment obligations and licence conditions.

Traditional dispute resolution usually requires negotiations, mediation, arbitration or court proceedings. An autonomous system can make this process faster by collecting evidence, identifying contractual rules, calculating amounts and recommending possible solutions. However, a machine should not automatically replace legally authorised human decision-makers where the law requires mediation, arbitration, administrative action or judicial review.

2. Meaning and Operation

An autonomous dispute-resolution system can operate through several stages:

Complaint registration – a party submits a dispute electronically.

Evidence collection – the system gathers contracts, meter readings, invoices, market records and communications.

Rule identification – software identifies the relevant legislation, licence condition, market rule or contractual provision.

Automated analysis – AI compares the evidence with the applicable rules.

Settlement recommendation – the system proposes possible solutions.

Mediation or arbitration support – the parties or authorised decision-maker use the analysis to resolve the dispute.

Review and appeal – parties retain legally available remedies.

NERSA's electricity division already has a Dispute Resolution and Customer Education function and states that it resolves disputes between industry participants, licensees and customers through mediation and arbitration.

3. Legal Framework in South Africa

The Electricity Regulation Act 4 of 2006 (ERA) provides an important statutory framework for electricity disputes. Section 30 deals with complaints and disputes involving electricity licensees and customers.

NERSA's existing dispute-resolution framework shows that electricity disputes can be handled through institutional processes rather than immediately going to court.

More recent draft NERSA Electricity Trading Rules also contemplate a staged process: parties should first attempt to resolve complaints themselves and, if unresolved, matters may be referred to NERSA for mediation, arbitration or investigation.

An autonomous system could support each stage, but it must operate within these legally established procedures.

4. Relevant Case Laws

Linde N.O. v Eskom Holdings SOC Ltd (2025)

In Linde N.O. and Others v Eskom Holdings SOC Ltd, the Free State High Court considered the relationship between Eskom's contractual dispute-resolution mechanism and the dispute-resolution process under section 30 of the ERA. The court noted that Eskom's standard supply conditions contained a dispute-resolution process, while section 30 also provided a statutory process.

This case is particularly relevant to autonomous systems because software must correctly identify which dispute-resolution mechanism applies. An algorithm cannot simply choose a procedure without considering the governing contract and legislation.

Eskom Holdings v Letsemeng Local Municipality (2022)

In Eskom Holdings SOC Ltd v Letsemeng Local Municipality, the dispute concerned substantial electricity debt and Eskom's proposed interruption of supply. The municipality sought to rely on the statutory dispute-resolution framework involving NERSA.

The case demonstrates that electricity disputes can have immediate consequences for supply. An autonomous system dealing with such disputes must therefore distinguish between ordinary commercial disagreement and urgent matters affecting electricity access or grid operations.

Eskom Holdings v Sonae Arauco (2024)

In Eskom Holdings v Sonae Arauco, the Supreme Court of Appeal considered a dispute involving loadshedding, a controlled-curtailment agreement and the statutory electricity codes governing grid protection. The court recognised Eskom's obligations concerning protection of the national grid.

This demonstrates that an automated dispute-resolution platform must consider not only private contractual rights but also public regulatory duties and grid-security requirements.

Mpact v Ekurhuleni Metropolitan Municipality (2026)

In Mpact (Pty) Ltd v Ekurhuleni Metropolitan Municipality, the High Court considered a dispute concerning electricity charges, municipal credit-control powers and a challenge involving NERSA-approved tariffs. The judgment discussed the interaction between electricity regulation, administrative law and constitutional principles.

The case illustrates why autonomous dispute-resolution systems must distinguish between private billing disputes and challenges to regulatory decisions. A machine cannot treat every dispute as an ordinary contractual disagreement.

5. Advantages of Autonomous Dispute Resolution

Faster Resolution

Automated systems can examine large amounts of information much faster than manual processes.

Lower Costs

Digital dispute platforms can reduce administrative expenses, particularly for routine billing or settlement disputes.

Better Evidence Management

Smart meters, transaction records and blockchain records can provide detailed evidence.

Early Settlement

AI can identify areas of agreement and suggest settlement options before a dispute becomes a formal proceeding.

Consistency

A properly designed system can apply the same predefined rules to similar disputes.

6. Major Legal Challenges

Due process is the first major issue. Parties must have a meaningful opportunity to present evidence and challenge adverse findings.

Transparency is also important. Parties should understand how an algorithm reached its recommendation.

Bias and errors can occur if the underlying data are inaccurate.

Human accountability is essential. An energy company or regulator should not avoid responsibility by saying that an automated system produced the result.

Confidentiality and cybersecurity are also important because energy disputes may involve commercially sensitive information.

Finally, enforceability must be considered. An AI-generated recommendation is not automatically a legally binding judgment or arbitral award.

7. Conclusion

Autonomous Dispute Resolution Systems can modernise energy-market dispute resolution by combining AI, digital evidence, automated analysis, mediation tools and smart-contract technology.

South African law already provides institutional mechanisms through NERSA for mediation, arbitration, complaints and investigations. Cases such as Linde v Eskom, Eskom v Letsemeng, Eskom v Sonae Arauco and Mpact v Ekurhuleni Municipality demonstrate that electricity disputes can involve overlapping contractual, regulatory, administrative and constitutional issues.

Therefore, the appropriate legal model is “automated assistance with legally accountable dispute resolution.” Autonomous technology can register complaints, analyse evidence, calculate claims and recommend settlements, but final legally binding decisions should remain subject to the applicable statutory procedure, contractual mechanism, arbitration framework or court process.