Topics · 02

Autonomous Agents, A2A Protocols & Platform Governance

Clarify legal status among clients, service providers and open platforms, reviewing delegation chains, Agent Card disclosures, task routing and high-impact action guardrails.

How do authority and legal liability travel across autonomous agents in multi-agent workflows?

Key Trigger Scenarios

Personal assistants calling third-party agents via the A2A protocol
Agent platforms publishing Agent Cards or admitting external autonomous agents
Agents authorized to execute financial transactions, system changes or communications
Multi-agent workflows transmitting personal data, trade secrets or transaction orders

Preliminary Due Diligence & Materials

01Map multi-stakeholder legal relationships between clients, providers, platforms and users
02Verify Agent Card truthfulness, API specifications and cryptographic authentication
03Separate ordinary automated tasks from high-impact actions requiring human confirmation
04Preserve end-to-end execution logs, exception triggers and manual override records

Selected Cases

Judicial Trends & Regulatory Standards

Read each case in its procedural context. Related cases may offer comparisons across technologies; they do not establish a single rule for every system.

Canada

Air Canada Chatbot Case

An airline website chatbot provided inaccurate bereavement fare information. The tribunal held that the enterprise is responsible for information provided on its website.

Tribunal Decision

The tribunal held that Air Canada owed a duty to take reasonable care over information published through its website chatbot and that the inaccurate answer amounted to negligent misrepresentation.
Air Canada was ordered to pay CAD 650.88 in damages, CAD 36.14 in pre-judgment interest, and CAD 125 in tribunal fees, for a total of CAD 812.02.

View Case Study →Singapore

Singapore Algorithmic Trading Contract Case

Cryptocurrency transactions were executed automatically by deterministic algorithms programmed by the parties at approximately 250 times the prevailing market price. The Court of Appeal held that valid contracts were formed by algorithmic execution and assessed unilateral mistake based on the programmer's state of mind at the time of coding.

Final Judgment

The Court of Appeal confirmed that algorithmic trading can form legally binding contracts.
The doctrine of unilateral mistake focuses on the knowledge and intention of the programmer when writing the underlying algorithm.

View Case Study →Colombia

Colombia Judicial Use of ChatGPT Case

A lower-court judge included ChatGPT responses in the reasoning of a judgment. The Constitutional Court found that the judge had independently decided the matter prior to querying the model, upholding the validity of the proceedings while establishing standards for judicial AI use including transparency, verification, privacy, and human control.

Constitutional Court Judgment

The Court held that generative AI did not displace the judge's independent decision-making in the specific proceeding.
The Court observed deficiencies in transparency and accountability regarding the manner in which the tool was used.

View Case Study →

Related Services

Regulatory Frameworks