1. Introduction

Autonomous financial market governance refers to the legal and institutional frameworks governing financial markets in which software, algorithms, artificial intelligence (AI), automated trading platforms and other autonomous systems perform functions traditionally undertaken by human traders, investment managers, brokers and compliance officers.

These systems may independently determine prices, generate orders, allocate investments, manage portfolios, detect suspicious transactions and respond to changing market conditions. Their increasing use creates significant questions of civil liability, contractual responsibility, investor protection, market integrity and regulatory accountability.

From a civil-law perspective, the central question is:

Who bears legal responsibility when an autonomous financial system causes loss, manipulates a market, breaches an investment mandate or makes a decision that violates financial regulations?

The answer depends on the relevant jurisdiction, the legal status of the parties, the contractual arrangements, the applicable regulatory duties and the circumstances of the loss.

This topic is best understood through a comparison of four major governance models:

Australia's statutory and common-law regulatory model.

The European Union's harmonised financial-market regulatory model.

The United States' enforcement-oriented securities regulatory model.

Civil-law jurisdictions such as Germany and France, where codified private law operates alongside specialised financial regulation.

These models share objectives but differ in how they allocate responsibility, supervise automated decisions and provide remedies to investors.

2. Meaning and scope of autonomous financial market governance

Autonomous financial market governance comprises the legal rules, supervisory mechanisms and internal controls governing automated financial decision-making.

It encompasses several categories of activity.

A. Algorithmic and high-frequency trading

Software determines order timing, price, quantity or execution strategy, potentially submitting and cancelling orders at very high speeds.

B. Autonomous investment management

Robo-advisers and portfolio algorithms select assets, rebalance investments and implement investment strategies based on client profiles or programmed objectives.

C. Automated compliance and surveillance

Systems monitor transactions, identify unusual trading patterns, flag suspected market abuse and generate regulatory reports.

D. Autonomous market infrastructure

Trading venues, clearing systems and settlement platforms use automated processes to match orders, manage collateral and process transactions.

The legal significance of autonomy lies in the extent to which a system operates without contemporaneous human intervention. It does not automatically mean that the software is legally independent or that responsibility transfers from the people and entities operating it.

3. Principal legal foundations

A. Civil liability

Civil liability concerns compensation or other remedies for loss caused by a legally actionable act or omission.

Potential causes of action include:

Negligence in designing, testing, deploying or supervising a trading system.

Breach of contract or an investment-management mandate.

Breach of fiduciary duties owed by an adviser or investment manager.

Misrepresentation concerning a system's capabilities, risks or performance.

Statutory liability for market manipulation, misleading conduct or regulatory breaches where the relevant legislation permits civil proceedings or remedies.

Liability arising from cybersecurity failures, unauthorised trading or defective market infrastructure.

The precise requirements differ by jurisdiction. A claimant generally must establish the elements of the relevant cause of action, including causation and legally recognised loss where required.

B. Contractual responsibility

Contracts are central to autonomous financial governance because they allocate duties between investors, brokers, exchanges, software vendors, custodians and investment managers.

Relevant contractual terms may address:

The permitted investment strategy and risk limits.

The authority to place and cancel orders.

Testing, maintenance and software updates.

Cybersecurity and access controls.

Audit rights and incident reporting.

Indemnities, liability limitations and dispute resolution.

A contract cannot necessarily exclude mandatory statutory duties or liability for conduct that the law does not permit parties to contract out of.

C. Regulatory responsibility

Financial regulation imposes obligations that may apply independently of a private contract. Depending on the activity, regulated firms may be required to maintain adequate systems, supervise personnel, manage conflicts of interest, preserve transaction records and monitor algorithmic trading.

The use of an external software provider does not automatically relieve a regulated firm of its own obligations.

D. Market integrity

Market integrity rules address conduct such as spoofing, manipulative order placement, misleading signals, insider dealing and other forms of prohibited market abuse.

An algorithm may execute a strategy automatically, but the legal analysis still examines the relevant conduct, the applicable statutory elements and the responsibility of the persons or entities involved.

4. Comparative governance models

Model I: Australia — statutory supervision and institutional accountability

Australia combines financial-services legislation, market-integrity rules, licensing requirements and judicial enforcement. The Australian Securities and Investments Commission (ASIC) supervises financial services and markets, while the Australian Securities Exchange (ASX) and other market operators have their own regulatory and operational responsibilities.

Important legal instruments include the Corporations Act 2001 (Cth), ASIC Act 2001 (Cth) and applicable market integrity rules.

The model emphasises:

Licensing and supervision of financial-service providers.

Obligations to provide services efficiently, honestly and fairly.

Market-manipulation and misleading-conduct prohibitions.

Adequate compliance systems, transaction monitoring and recordkeeping.

Civil penalties, injunctions and other statutory remedies.

An autonomous trading platform is not itself a substitute for the licensed entity's legal responsibilities. A financial firm that deploys an algorithm must assess its legal obligations, including applicable controls, supervision and incident management.

A useful illustration is Australian Securities and Investments Commission v Kobelt [2019] HCA 18. Although the case did not concern autonomous trading, it demonstrates the role of statutory consumer protection and the difficulty of applying an unconscionability standard to financial arrangements. The High Court, by majority, rejected ASIC's appeal concerning unconscionability, while the separate credit-licensing issue had been resolved against the respondent below.

Model II: European Union — harmonised algorithmic-market regulation

The EU model combines legislation applicable across member states with supervision by national competent authorities and coordination by the European Securities and Markets Authority (ESMA).

The principal instruments include:

Markets in Financial Instruments Directive II (MiFID II).

Markets in Financial Instruments Regulation (MiFIR).

Market Abuse Regulation (MAR).

Relevant delegated regulations and regulatory technical standards.

MiFID II addresses algorithmic trading, including systems that automatically determine order parameters with limited or no human intervention. Regulatory requirements address testing, risk controls, monitoring, business continuity and the prevention of disorderly markets.

ESMA guidance also makes clear that a firm remains responsible for compliance when it uses a third-party system with algorithmic trading functionality. Outsourcing technology does not necessarily outsource the firm's regulatory responsibility.

The EU model's main strength is its detailed treatment of automated trading as a regulated activity rather than merely a technical feature. Its challenges include cross-border supervision, differences in enforcement practice and the difficulty of supervising rapidly changing algorithms.

Model III: United States — enforcement and market-integrity controls

The United States relies on federal securities and commodities laws, specialised regulators and extensive enforcement powers.

The Securities and Exchange Commission (SEC) supervises securities markets and investment advisers, while the Commodity Futures Trading Commission (CFTC) oversees relevant derivatives markets.

Important legislation includes the Securities Exchange Act of 1934 and the Commodity Exchange Act.

The US approach addresses:

Manipulative orders and spoofing.

Fraudulent representations and deceptive trading practices.

Broker-dealer and investment-adviser obligations.

Exchange and clearing infrastructure risks.

Civil enforcement, administrative proceedings and criminal prosecution where authorised.

A key characteristic is the use of enforcement proceedings to establish how established market-abuse laws apply to new technological methods. This provides flexibility, although firms may face uncertainty where legal standards must be applied to unfamiliar automated strategies.

Model IV: Germany and France — codified duties and regulated institutions

Germany and France provide useful civil-law comparisons because their private-law systems are organised around codified legal principles while their financial markets are also subject to extensive specialised legislation.

In Germany, the Civil Code (Bürgerliches Gesetzbuch, BGB) provides general principles concerning contractual obligations and damages, while securities and financial-market legislation governs regulated activities and market conduct.

In France, the Civil Code governs general contractual liability, while the Monetary and Financial Code and securities-market rules establish financial-sector requirements. The Autorité des marchés financiers (AMF) supervises relevant French financial markets.

These jurisdictions demonstrate that civil-law codification is compatible with detailed algorithmic trading regulation, specialised supervision and judicial remedies.

The main comparative distinction is structural: civil-law private-law rules may be stated in general codified provisions, while sector-specific legislation defines the regulatory duties of financial institutions. Neither structure eliminates the need to determine who controlled a system, what duties applied and whether a breach caused the loss.

5. Comparison of the four models

The models are not mutually exclusive. In practice, a multinational financial institution may be subject to several of them simultaneously.

6. Case laws on autonomous financial market governance

The following cases are particularly useful for studying responsibility for automated trading, market manipulation, institutional compliance and financial-service conduct. An important qualification is that most are not AI-liability cases in the narrow sense. They establish legal principles that may apply when autonomous systems perform regulated financial activities.

Case 1: SEC v. Lek Securities Corp. and others (S.D.N.Y. 2019; Second Circuit 2022)

Jurisdiction: United States.

Facts: The SEC pursued a case involving a brokerage firm and individuals associated with trading schemes involving layering and cross-market manipulation. The schemes used orders and cancellations to create misleading market signals and influence securities and options prices.

Legal issue: When can traders and intermediaries be held liable for facilitating manipulative trading conducted through sophisticated electronic trading arrangements?

Decision: A jury found the principal trading defendants liable for market manipulation in 2019. In 2022, the Second Circuit affirmed the judgment against the relevant trading defendants.

Legal principle: Electronic execution does not shield a trading strategy from market-manipulation prohibitions. The legal analysis focuses on the conduct, the applicable statutory elements and the responsibility of the participants.

Relevance to autonomous governance: Firms operating algorithmic trading systems need controls to detect layering, suspicious order cancellations and manipulative interactions across markets. Brokers and intermediaries may face liability where their own conduct satisfies the applicable legal requirements.

Case 2: CFTC v. Navinder Singh Sarao and Nav Sarao Futures Limited PLC (N.D. Ill. 2016)

Jurisdiction: United States.

Facts: Sarao used an automated trading program and other methods to place large orders in E-mini S&P 500 futures without intending that the orders be executed. The orders created misleading impressions of supply and demand. The CFTC alleged that the activity contributed to market conditions surrounding the May 2010 Flash Crash.

Legal issue: Can automated trading strategies that create artificial market signals constitute unlawful spoofing and manipulation?

Decision: In 2016, the federal court entered a consent order requiring monetary sanctions, disgorgement and trading restrictions. Sarao admitted the relevant allegations in the order.

Legal principle: Automation does not neutralise liability for prohibited market conduct. A trading program can be an instrument of manipulation, and the person or entity responsible may face civil enforcement.

Relevance to autonomous governance: This case illustrates why financial institutions need order-level monitoring, cancellation-pattern analysis, risk limits and the capacity to suspend trading strategies that generate misleading market signals.

Case 3: CFTC v. Kraft Foods Group, Inc. and Mondelēz Global LLC (N.D. Ill.; consent order in 2022)

Jurisdiction: United States.

Facts: The CFTC alleged that the companies pursued a strategy involving substantial wheat-futures positions in circumstances intended to affect the relationship between futures prices and cash-wheat prices. The proceedings also involved allegations concerning speculative position limits and noncompetitive transactions.

Legal issue: When does trading intended to influence a related market price constitute unlawful manipulation?

Outcome: The litigation culminated in a consent order in 2022 resolving the CFTC's claims.

Legal principle: The legality of a trading strategy depends on the governing commodities legislation and the relevant evidence concerning manipulation, attempted manipulation and other alleged violations.

Relevance to autonomous governance: An algorithm may implement a strategy consistently and efficiently, but automation does not make a prohibited market strategy lawful. Firms must evaluate both the technical execution and the legal purpose and effects of their trading activity.

Case 4: Australian Securities and Investments Commission v Commonwealth Securities Limited and Australian Investment Exchange Limited (Federal Court of Australia, 2022)

Jurisdiction: Australia.

Facts: CommSec and AUSIEX were subject to proceedings concerning systemic compliance failures, including brokerage overcharging, deficiencies in transaction confirmations, best-execution compliance and systems for identifying possible wash trading.

Legal issue: What responsibility do financial-services firms bear for maintaining systems and processes that comply with market-integrity rules and statutory obligations?

Outcome: In October 2022, the Federal Court ordered penalties of $20 million against CommSec and $7.12 million against AUSIEX.

Legal principle: Financial institutions must maintain adequate compliance systems and processes. The fact that a business operates through electronic platforms does not eliminate its obligations to customers or the market.

Relevance to autonomous governance: This is a particularly useful Australian example of institutional accountability for trading infrastructure and compliance controls. It supports the proposition that technology-enabled services must be governed by systems capable of delivering accurate transactions, fair execution and regulatory compliance.

Case 5: Australian Securities and Investments Commission v Kobelt [2019] HCA 18

Jurisdiction: Australia.

Facts: The case concerned a store operator's book-up credit system, under which customers supplied bank cards and PINs and the operator controlled withdrawals to recover credit. ASIC alleged that the system involved unconscionable conduct and breaches of credit legislation.

Legal issue: How should statutory protections against unconscionable financial conduct be applied to a particular financial arrangement?

Decision: The High Court, by a majority of four to three, dismissed ASIC's appeal on unconscionability. The separate credit-licensing contravention remained established.

Legal principle: Financial conduct must be assessed under the applicable statutory standard and in light of the circumstances of the arrangement. Unconscionability cannot simply be presumed from the existence of an unusual or disadvantageous financial arrangement.

Relevance to autonomous governance: Although not an algorithmic trading case, Kobelt illustrates the importance of statutory consumer protections, factual context and legal accountability when financial systems affect customers' control over their money.

Case 6: Spector Photo Group NV v Commissie voor het Bank-, Financie- en Assurantiewezen (Case C-45/08, 2009)

Jurisdiction: European Union.

Facts: The case concerned insider dealing and the interpretation of EU market-abuse rules, including the relationship between possession of inside information and trading in financial instruments.

Legal issue: How should EU market-abuse legislation be interpreted when determining whether a person who possesses inside information has engaged in prohibited insider dealing?

Decision: The Court of Justice of the European Union considered the statutory framework for insider dealing and the operation of the relevant evidential presumptions.

Legal principle: Market-abuse rules must be interpreted in accordance with their statutory conditions and the legal safeguards governing the determination of prohibited conduct.

Relevance to autonomous governance: Where an automated trading system receives or processes sensitive information, the legal question is whether the relevant persons and entities have breached applicable insider-dealing rules. The use of software does not displace the need to establish the statutory elements of liability.

Case 7: Geltl v Daimler AG (Case C-19/11, 2012)

Jurisdiction: European Union.

Facts: The case involved disclosure obligations relating to inside information and intermediate steps in a process leading to a significant corporate event.

Legal issue: When can an intermediate step in a longer process constitute inside information requiring disclosure under EU market-abuse legislation?

Decision: The Court of Justice examined the meaning of information that is sufficiently precise and the treatment of intermediate steps in a protracted process.

Legal principle: Disclosure obligations must be assessed by reference to the applicable legal definition of inside information, including the statutory criteria concerning precision and price sensitivity.

Relevance to autonomous governance: Automated surveillance and disclosure systems must correctly distinguish potentially material information from ordinary market data. Firms remain responsible for ensuring that automated processes comply with applicable disclosure obligations.

Case 8: ASIC v Westpac Securities Administration Ltd [2021] FCAFC 55

Jurisdiction: Australia.

Facts: The proceedings concerned Westpac's conduct in relation to superannuation and financial advice, including questions concerning statutory obligations and the treatment of customers' financial interests.

Legal issue: How should financial-services obligations be applied to institutional conduct affecting customers' financial outcomes?

Decision: The Full Court of the Federal Court considered the relevant statutory obligations and the legal character of Westpac's conduct.

Legal principle: Financial institutions must comply with the duties imposed by the applicable financial-services legislation; the commercial structure of a service does not itself determine whether those duties have been met.

Relevance to autonomous governance: The case provides a broader institutional-accountability perspective. In automated financial services, firms must assess whether the design and operation of a system satisfy the obligations applicable to the service being provided.

Comparative lessons from the cases

7. Civil liability when autonomous financial systems cause loss

Civil liability is one of the most important aspects of autonomous financial governance because an algorithmic error can affect numerous investors within seconds.

A. Negligence

A claim in negligence may arise when a financial institution fails to take reasonable care in designing, testing, deploying or supervising an automated system.

Potential examples include:

Deploying an algorithm without appropriate testing.

Failing to impose adequate trading limits.

Ignoring known software defects.

Failing to respond to repeated system alerts.

Allowing unauthorised access to a trading platform.

A claimant must establish the elements required by the applicable law, including duty, breach, causation and legally recoverable damage. A system malfunction does not automatically establish negligence.

B. Breach of contract

A financial institution may be contractually liable if its automated system fails to perform an agreed service in accordance with the contract.

For example, an investment manager may agree to operate a portfolio within specified risk limits. If an algorithm repeatedly exceeds those limits, the resulting conduct may constitute a breach of contract, depending on the contractual terms and surrounding circumstances.

The court must examine the agreement, the scope of authority granted to the manager and any applicable statutory protections.

C. Fiduciary duties

Investment managers and other fiduciaries may owe duties concerning loyalty, conflicts of interest and the proper exercise of entrusted powers.

An autonomous system does not remove those duties. A manager that delegates investment decisions to software must consider whether the delegation is legally permissible, whether conflicts are controlled and whether the manager has discharged the duties that remain applicable.

D. Product and software liability

Defective financial software may raise questions concerning contractual warranties, negligence, statutory consumer protections or product liability, depending on the product and jurisdiction.

Potential defendants may include developers, vendors, financial institutions and platform operators. Responsibility is not automatically shared equally: it depends on each party's role, duties, conduct and the causal connection with the loss.

E. Cybersecurity and unauthorised transactions

Where a system is compromised, liability may depend on the adequacy of security controls, contractual allocation of risk, statutory obligations and the conduct of the parties.

Relevant issues include authentication, access permissions, transaction verification, incident reporting, recovery procedures and the preservation of audit records.

8. A comparative framework for allocating responsibility

A useful method is to identify the entity responsible for each stage of the automated financial process.

1. System design and development

Software developer, vendor or internal technology team

2. Deployment and risk approval

Financial institution, investment manager or trading firm

3. Execution and market access

Trader, broker, exchange or electronic platform operator

4. Monitoring and incident response

Compliance teams, risk officers and responsible management

5. Legal review and remedies

Financial regulators, courts and dispute-resolution bodies

This framework is analytical rather than a universal legal rule. One entity may perform several functions, and liability may arise at multiple stages. The purpose is to identify the relevant duty, breach, causal contribution and available remedy.

9. Regulatory safeguards for autonomous markets

Effective governance requires controls before, during and after automated trading.

The EU framework expressly addresses governance of trading venues and algorithmic trading. Australia and the United States also impose relevant obligations through financial-services, market-integrity and market-abuse laws.

The precise controls required depend on the type of firm, trading activity, jurisdiction and applicable rules.

10. Principal challenges and future developments

A. The accountability gap

A complex trading outcome may result from the interaction of multiple systems, data feeds, models and market conditions. It can be difficult to determine whether a loss arose from faulty design, inadequate supervision, misleading inputs or a market event.

Legal frameworks should preserve clear accountability without automatically treating every adverse outcome as proof of wrongdoing.

B. Explainability and evidence

Regulators and courts may need to understand why an algorithm placed an order, what information it used and whether its operation complied with relevant rules.

Black-box systems can complicate this inquiry. Reliable records, documented testing, version control and access to relevant technical evidence are therefore important.

C. Market concentration and systemic risk

Similar algorithms may respond to the same signals and amplify volatility. Automated trading can also create dependencies between market participants and critical infrastructure.

Effective governance must consider not only individual misconduct but also the possibility of widespread disruption.

D. Cross-border enforcement

A single transaction may involve a software provider in one country, a broker in another and an exchange elsewhere. The applicable laws, regulatory jurisdiction, evidence-gathering powers and available remedies may differ.

International cooperation and consistent standards are therefore important to effective oversight.

E. Fairness and investor protection

Automated systems may provide efficiency and lower transaction costs, but they can also create information asymmetries and expose investors to risks they do not understand.

Regulators must balance innovation with transparency, appropriate risk disclosure and protection against unlawful or unfair conduct.

11. Overall conclusion

Comparative autonomous financial market governance demonstrates that financial regulation must adapt to systems capable of making and executing decisions at extraordinary speed without eliminating established principles of civil liability and institutional responsibility.

Australia emphasises statutory supervision and accountability of licensed financial institutions. The European Union provides a detailed harmonised framework for algorithmic trading. The United States has developed extensive enforcement mechanisms addressing manipulation and deceptive trading. Germany and France demonstrate how codified private law can operate alongside sophisticated financial-market regulation.

The case law, particularly SEC v. Lek Securities, CFTC v. Sarao, ASIC v. CommSec and AUSIEX, and ASIC v. Kobelt, illustrates that automated systems must operate within existing legal requirements governing market conduct, financial services and consumer protection. These decisions should not be mistaken for a general judicial rule that all algorithmic losses create liability.

The central principle is that autonomous technology changes how financial decisions are made, but does not by itself eliminate the legal responsibilities of the institutions and individuals who design, deploy, control or use it.

A sound governance model must therefore combine technical safeguards, transparent decision-making, effective supervision, enforceable legal duties and accessible remedies for those harmed by unlawful financial conduct.