1. Introduction

The increasing use of artificial intelligence (AI), machine learning, automated bidding, predictive analytics and algorithmic trading is transforming electricity and other energy markets. Traditionally, market participants—generators, traders, distribution companies and consumers—made bidding and purchasing decisions through human judgment. AI-driven markets increasingly allow software systems to forecast demand, predict renewable generation, determine bids, optimise storage and automatically execute transactions.

An AI-driven energy market can therefore be understood as an energy market in which algorithms or AI systems materially influence price discovery, bidding, dispatch, trading, congestion management, demand response, energy storage and market surveillance.

This creates a distinctive governance problem. Electricity is not an ordinary commodity: supply and demand must be balanced almost instantaneously, transmission networks are physically constrained, and market power can produce substantial consumer effects. Consequently, AI governance must combine energy regulation, competition law, market-abuse rules, data governance, cybersecurity, consumer protection and administrative law.

The issue is becoming particularly important in Europe, where ACER is actively examining algorithmic trading under the EU's wholesale-energy market-integrity framework. (ACER)

2. Meaning and Characteristics of AI-Driven Energy Markets

AI-driven energy markets involve several technologies:

Machine-learning bidding systems – algorithms determine electricity bids based on expected prices and system conditions.

Demand forecasting – AI predicts electricity consumption.

Renewable forecasting – algorithms forecast wind and solar generation.

Automated trading – software submits and modifies market orders without continuous human intervention.

Battery optimisation – AI decides when storage should charge or discharge.

Demand-response algorithms – AI automatically changes consumption according to price or grid conditions.

Congestion management – algorithms optimise electricity flows through constrained networks.

Market surveillance – AI identifies suspicious bidding patterns or possible manipulation.

Predictive maintenance – AI predicts failures in generators, transformers and transmission equipment.

The EU's revised REMIT framework expressly recognises algorithmic trading in wholesale energy products, including systems that automatically determine whether to initiate an order and its timing, price or quantity. (EUR-Lex)

3. Why AI Creates a New Governance Problem

AI can improve efficiency, but it also changes the relationship between human decision-makers and market outcomes.

A traditional market participant can generally explain:

"We submitted this bid because our marginal cost was X."

An AI system may instead generate a bid because thousands of historical and real-time variables produce a particular prediction.

This raises questions such as:

Who is legally responsible for an AI-generated bid?

Must an algorithm be explainable?

Can an AI system manipulate prices without being explicitly programmed to manipulate?

Who bears liability when an automated trading system causes a grid disturbance?

Should regulators have access to training data?

How should regulators audit proprietary algorithms?

Can competing AI systems independently learn behaviour resembling collusion?

Recent research specifically identifies the possibility that learning-based agents in electricity markets could develop tacit-collusive outcomes without being explicitly instructed to collude. (arXiv)

Thus, traditional regulatory concepts such as intent, control, knowledge and responsibility become more complicated.

4. Legal Foundations of Governance

AI-driven energy markets require a multi-layered regulatory structure.

A. Electricity law

Electricity regulators retain authority over:

market design;

licensing;

trading;

transmission;

system operation;

balancing;

tariffs;

ancillary services;

reliability.

In India, the Electricity Act 2003 provides the basic statutory framework, while CERC's Power Market Regulations regulate electricity-market institutions and transactions.

The CERC Power Market Regulations 2021 expressly contemplate automated audit trails, requiring electronic time-sequenced records of transactions and their creation, modification and deletion. (CERC)

This is particularly relevant to AI because an automated audit trail can establish what an algorithm actually did even where the decision-making process is highly complex.

5. Governance Through Transparency

Transparency is one of the most important principles for AI-driven energy markets.

A regulator should potentially be able to determine:

which algorithm submitted a bid;

when the algorithm acted;

what data it received;

whether a human approved the transaction;

whether the algorithm changed its strategy;

what orders were cancelled;

whether abnormal behaviour occurred.

The European Union has strengthened wholesale-energy transparency through REMIT. In 2026, the European Commission adopted new rules concerning energy-market data reporting and supervision, designed to improve ACER's ability to monitor wholesale markets and detect market abuse. (Energy)

Indian position

The CERC regulatory framework similarly demonstrates the importance of transaction records and market surveillance. CERC has historically exercised direct oversight over power-exchange software and algorithms.

Notably, CERC records include a 2011 suo-motu proceeding concerning the audit of trading-software algorithms used for price discovery by Indian power exchanges. (CERC)

This is an important early example of the principle that the software used to determine electricity-market outcomes can itself become a regulatory object.

6. Algorithmic Price Discovery

Price discovery is central to electricity-market governance.

An AI system may consider:

anticipated demand;

generator availability;

fuel prices;

weather;

renewable output;

transmission congestion;

storage levels;

competing bids;

historical prices.

The danger is that an algorithm may systematically produce prices that are inconsistent with competitive market behaviour.

CERC has previously scrutinised the design of electricity-exchange price-discovery mechanisms.

Case: Power Exchange India Ltd. In Re

CERC examined proposed price-discovery and matching mechanisms and expressed concerns about information asymmetry, unequal opportunities and effects on competition between power exchanges. The Commission modified the proposed mechanism rather than allowing the exchange unrestricted control over price discovery. (CaseMine)

The principle is highly relevant to AI governance:

Market participants may innovate in trading technology, but the fundamental architecture of electricity price discovery remains subject to regulatory oversight.

7. Case Law: PTC India Ltd. v. CERC

The Supreme Court's decision in PTC India Ltd. v. Central Electricity Regulatory Commission, (2010) 4 SCC 603 is foundational for understanding electricity-sector regulation in India. (Indian Kanoon)

The Court recognised the extensive regulatory role of CERC under the Electricity Act 2003 and distinguished regulatory functions from ordinary adjudication.

Importance for AI markets

AI systems do not displace statutory regulatory authority.

Even if:

an exchange uses sophisticated algorithms;

generators employ autonomous bidding;

traders use machine-learning systems;

the ultimate market remains subject to the statutory framework established by Parliament and implemented by the electricity regulator.

Therefore, AI cannot be treated as an independent legal authority capable of replacing CERC's regulatory jurisdiction.

8. Case Law: Indian Energy Exchange Ltd. v. CERC

In Indian Energy Exchange Ltd. v. CERC, the Appellate Tribunal for Electricity considered regulatory requirements imposed upon the power exchange concerning its professional members and handling of client funds. (Indian Kanoon)

The case demonstrates a broader principle:

Power exchanges are regulated market institutions, rather than purely private technological platforms.

This becomes even more significant when exchanges employ AI.

An AI-driven exchange cannot argue that an algorithm is merely a technological tool and therefore outside regulatory supervision. The underlying activity—electricity trading—remains regulated.

9. Case Law: Power Exchange India Ltd. v. POSOCO

In Power Exchange India Ltd. v. Power System Operation Corporation Ltd., CERC considered transmission-corridor allocation and competition between power exchanges. (CaseMine)

The Commission considered:

market shares;

congestion;

transmission constraints;

competition between exchanges;

market coupling;

equitable access.

The case is important for AI governance because AI-based trading cannot be separated from physical grid constraints.

An algorithm may optimise a trading strategy perfectly from a financial perspective while simultaneously producing excessive transactions across a congested transmission corridor.

Consequently:

AI market optimisation must remain subordinate to physical-system security and regulated network constraints.

10. Competition Law and AI-Driven Energy Markets

Competition law is another major pillar.

AI can potentially produce:

Pro-competitive effects

better forecasting;

lower transaction costs;

greater market liquidity;

more efficient dispatch;

reduced balancing costs;

improved renewable integration.

Anti-competitive risks

algorithmic price coordination;

tacit collusion;

discriminatory access;

exclusion of smaller traders;

manipulation of bids;

strategic withholding;

exploitation of market power.

The electricity sector is particularly sensitive because electricity markets frequently have limited suppliers, network bottlenecks and repeated interactions.

Research on AI-based electricity markets has identified the possibility of learning algorithms sustaining supra-competitive outcomes even without explicit instructions to collude. (arXiv)

11. The Problem of Tacit Algorithmic Collusion

Suppose Generator A and Generator B independently employ reinforcement-learning systems.

Neither company instructs its AI to collude.

However, the algorithms repeatedly observe:

Generator A's bids;

Generator B's bids;

market prices;

demand;

competitor responses.

Over time, both algorithms may learn that maintaining high prices is more profitable than aggressive competition.

The legal difficulty is that traditional cartel law generally looks for some form of agreement, coordination or concerted practice.

AI may create a situation in which:

human agreement → absent

but

algorithmic coordination → economically observable.

This creates a challenge for competition authorities.

Governance may therefore require regulators to monitor outcomes and behavioural patterns, not merely communications between human executives.

12. Market Manipulation

AI may also be used for:

spoofing;

layering;

artificial demand signals;

false bids;

strategic cancellations;

exploiting information asymmetry;

manipulating congestion;

creating artificial scarcity.

REMIT is particularly significant in Europe because it establishes an EU-wide framework for detecting and deterring manipulation and abuse in wholesale energy markets. ACER has specifically begun adapting its surveillance practices to algorithmic trading. (ACER)

Therefore, AI governance requires continuous market surveillance, rather than relying exclusively upon investigations after manipulation has occurred.

13. Accountability and Liability

One of the most difficult legal questions is:

Who is responsible when AI makes an unlawful decision?

Possible responsible parties include:

generator;

electricity trader;

power exchange;

software developer;

algorithm operator;

system operator;

corporate directors;

regulated licensee.

A sound governance framework should avoid creating an accountability gap.

The principle should be:

Deployment of autonomous technology should not eliminate legal responsibility.

A company cannot necessarily avoid regulatory responsibility merely by saying:

"The algorithm did it."

Human and corporate accountability must remain attached to the regulated activity.

14. Explainability

AI models can be highly complex, especially deep-learning systems.

Regulators may therefore require different degrees of explanation depending upon risk.

For example:

A regulator does not necessarily need access to every line of source code. It may instead require:

model documentation;

audit logs;

input/output records;

decision explanations;

testing results;

risk assessments;

incident reports.

15. Data Governance

AI-driven energy markets depend heavily on data.

Relevant information includes:

electricity consumption;

smart-meter information;

generator output;

market bids;

transmission capacity;

weather information;

consumer behaviour;

pricing information.

Energy regulators must balance:

data access + competition + privacy + cybersecurity.

Too little data may make AI systems inaccurate.

Too much unrestricted data may facilitate:

privacy violations;

market manipulation;

competitive intelligence;

cyberattacks.

Therefore, governance should establish clear rules regarding who may collect, process, share and retain energy-market data.

16. Cybersecurity

AI increases the attack surface of energy markets.

An attacker might attempt to:

manipulate training data;

corrupt forecasts;

alter bids;

interfere with automated dispatch;

create false congestion signals;

attack market exchanges;

compromise distributed energy resources.

A cyberattack against an AI-based energy-management system could therefore have both economic and physical consequences.

Governance should consequently require:

cybersecurity standards;

authentication;

access controls;

model monitoring;

incident reporting;

backup systems;

human override mechanisms.

17. Human Oversight

A central governance principle should be human-in-the-loop regulation.

Human oversight is especially important for:

emergency grid conditions;

automatic disconnection;

extreme price events;

market suspension;

cybersecurity incidents;

abnormal algorithmic behaviour.

AI can recommend or execute actions, but regulators and system operators should retain the legal authority to intervene.

This principle is particularly important because electricity systems can experience sudden physical events that fall outside historical training data.

18. AI and Market Coupling

Modern electricity markets increasingly use market coupling to coordinate multiple bidding zones.

CERC has considered market coupling in relation to Indian power exchanges, including directions concerning implementation of market coupling in 2026. (CERC)

AI could potentially improve:

cross-border optimisation;

congestion management;

transmission utilisation;

renewable integration;

balancing.

But centralised algorithmic market coupling also creates governance concerns because a malfunction could affect multiple markets simultaneously.

Therefore, market-coupling algorithms require:

independent testing;

transparent rules;

cybersecurity;

auditability;

contingency mechanisms.

19. AI and Consumer Protection

AI-driven wholesale markets ultimately affect electricity consumers.

AI could reduce costs through:

better demand forecasting;

efficient dispatch;

lower balancing costs;

improved renewable utilisation.

But algorithmic systems could also create:

unpredictable prices;

discriminatory pricing;

automated service restrictions;

complex tariffs;

difficulty challenging automated decisions.

Consumers therefore need:

understandable billing;

transparency regarding automated decisions;

complaint mechanisms;

protection against discriminatory outcomes;

regulatory oversight of automated pricing.

20. Governance Model for AI-Driven Energy Markets

A comprehensive governance framework can be organised into eight pillars:

Pillar 1 — Registration

High-impact energy algorithms should be identifiable to the regulator.

Pillar 2 — Risk classification

Algorithms should be classified according to their potential impact.

Pillar 3 — Testing

AI systems should undergo testing before deployment.

Pillar 4 — Auditability

Complete transaction and decision logs should be maintained.

Pillar 5 — Human oversight

Regulators and system operators should retain intervention powers.

Pillar 6 — Competition monitoring

Authorities should monitor algorithmic coordination and market power.

Pillar 7 — Cybersecurity

AI systems should comply with critical-infrastructure security requirements.

Pillar 8 — Liability

The use of AI should not eliminate responsibility of regulated entities.

21. Indian Legal Position

India does not yet have a single comprehensive statute specifically governing AI-driven electricity markets. Instead, governance is likely to arise from the interaction of:

Electricity Act 2003;

CERC regulations;

Power Market Regulations;

Grid Code;

competition law;

information-technology and cybersecurity rules;

consumer-protection principles;

contractual and corporate liability.

CERC's current regulatory framework continues to evolve. Its 2026 regulatory materials include amendments concerning power markets and work concerning market coupling and real-time-market arrangements. (CERC)

The existing framework therefore provides many of the necessary regulatory foundations, even though AI-specific governance remains an emerging area.

22. Important Case Laws and Their Relevance

The Tata Power litigation concerned important questions concerning electricity regulation, distribution and statutory market structure under the Electricity Act. (Indian Kanoon)

23. Key Legal Principles Emerging

Several principles can be derived from the existing regulatory framework and case law:

1. Technology neutrality

The law should regulate the economic activity and its consequences, rather than allowing AI to create a regulatory loophole.

2. Algorithmic accountability

A regulated company remains responsible for the systems it deploys.

3. Auditability

Automated market decisions should be reconstructable after the event.

4. Competition protection

AI should not facilitate market manipulation or anti-competitive coordination.

5. Grid security

Commercial optimisation must remain subordinate to electricity-system security.

6. Procedural fairness

Regulated entities should have meaningful opportunities to challenge regulatory decisions concerning automated systems.

7. Proportionality

The more consequential the AI system, the stronger the governance requirements should be.

24. Future Challenges

Future AI-driven energy markets will raise increasingly difficult legal questions concerning:

autonomous electricity traders;

AI-generated PPAs;

AI-controlled virtual power plants;

autonomous battery trading;

AI-managed microgrids;

machine-to-machine electricity contracts;

AI-based demand response;

algorithmic capacity markets;

AI-controlled cross-border trading;

blockchain and AI combinations;

autonomous energy communities.

The rapid expansion of AI infrastructure itself also creates additional electricity-demand and grid-planning issues. Recent European policy discussions, for example, have focused on the growing electricity requirements of AI data centres and the resulting pressure on grid infrastructure. (Reuters)

25. Conclusion

Governance of AI-driven energy markets is fundamentally a question of controlling the relationship between autonomous technology, market competition and physical electricity infrastructure.

AI can improve forecasting, trading, renewable integration, storage optimisation and system efficiency. However, autonomous algorithms can also amplify market power, create new forms of manipulation, generate tacit coordination and make responsibility more difficult to establish.

Indian electricity jurisprudence already provides important foundations. PTC India v. CERC establishes the significance of statutory regulatory authority; power-exchange decisions demonstrate that price discovery, competition and exchange operations remain subject to regulatory supervision; and CERC's existing Power Market Regulations provide mechanisms such as automated audit trails. (Indian Kanoon)

The future legal framework should therefore combine algorithmic transparency, auditability, competition monitoring, cybersecurity, human oversight, data governance and clear liability rules.

The central principle can be stated simply:

AI may automate decisions in an energy market, but automation should not automate away accountability, competition, transparency or regulatory control.

This makes AI-driven energy-market governance an emerging field at the intersection of Energy Law, Artificial Intelligence Law, Competition Law, Administrative Law, Cybersecurity Law and Regulatory Governance.