AI Agents
16 articles
What Is Discoverable When an AI Agent Is the Actor?
When the conduct at issue was performed by an autonomous agent, the evidence is not email and testimony. It is system prompts, tool-call logs, model versions, and retrieval context—most of it ephemeral by default. What to preserve, what to request, and why the preservation duty attaches earlier than teams expect.
Can a California Court Reach the Company Behind an AI Agent?
An AI agent transacts with a Californian and something goes wrong. The agent is not a legal person, so the defendant has to be the company that deployed it—and that company may sit anywhere. After Briskin and Gelasio, the borderless-internet defense is gone in the Ninth Circuit. Here is how the express-aiming test applies when the thing doing the aiming chose its own counterparty.
Know Your Agent (KYA): The Legal Compliance Standard for AI Agents That Move Money
AI agents can now initiate real payments, and every technical standard — Visa, Mastercard, Google, NIST — builds cryptographic identity without answering the only question the law cares about: which legal person answers for the transaction. Know Your Agent (KYA) is the compliance standard that answers it.
Who Is Liable When an AI Agent Loses Your Money?
AI agents now trade, pay, and move money on their own, and some of them have already lost it. No AI-liability statute governs the loss. Existing law does, and it keeps asking one question the industry cannot dodge: who controlled the agent?
Can an AI Agent Legally Enter Into a Contract?
A 1999 uniform act and a 2000 federal statute answered whether a machine can make a contract. UETA and ESIGN call the software an 'electronic agent' and say a contract it forms is not void just because no human reviewed it, where the agent's act is legally attributable to the person bound. That act is the deployer's own, ordinarily leaving the deployer bound, mistakes included, save at narrow edges in UETA and common law. The genuinely open questions are about authority and error, not validity.
Does Your Agentic-Payments Startup Need a Money Transmitter License?
When an AI agent moves money on a user's behalf, the licensing question turns on one thing regulators have asked for decades: do you control the funds? If your platform holds, pools, or controls customer money, you are likely a money transmitter — federally and in most states — no matter how autonomous the agent is.
Does Your AI Agent Need a Financial License? A Decision Guide
Whether an AI agent needs a financial license does not turn on the fact that it is AI. It turns on what the agent does with money or securities — and more than one regime can apply at once. An agent that executes securities trades answers to the SEC; one that trades futures, swaps, or leveraged retail crypto answers to the CFTC; one that moves customer money answers to FinCEN and the states. This is the decision guide that routes your agent to the right regulator — often more than one.
Who Is the BSA Customer When an AI Agent Sends Stablecoins?
The GENIUS Act made stablecoin issuers Bank Secrecy Act institutions, and AI agents are now sending stablecoins on their own. The customer-identification rules assume a person. Corporate law has answered the non-human-person question for two hundred years.
Does Your AI Trading Agent Need to Register as a Broker-Dealer?
When an AI agent executes trades for other people, the 1934 Act asks one thing: for whose account? An agent that effects transactions for someone else's account is a broker and generally must register; one that trades its own account may be an exempt 'trader,' and one that matches other users' orders can become an exchange. Autonomy does not change which hat you wear.
Does Your AI Trading Agent Need to Register with the CFTC?
The CFTC asks two questions the SEC's framework does not: what does the agent trade, and what job does it do with it? Futures, swaps, and leveraged retail crypto pull an autonomous trading agent into the Commodity Exchange Act — and depending on whether it advises, pools capital, or takes orders, it can owe registration as a CTA, CPO, FCM, or introducing broker. The CFTC's Ooki DAO enforcement action signals that being autonomous does not dissolve the duty.
Does Your AI Trading Agent Need to Register as an Investment Adviser?
When an AI agent recommends securities to clients for a fee, the Investment Advisers Act asks one thing: what is the advice, not who — or what — delivers it. An autonomous agent that gives personalized investment advice for compensation is judged by the same three-part test as a human adviser, and no line of code creates an exemption.
Drafting Vendor Agreements for Claude Managed Agents: A GC's Clause Library
Standard SaaS MSAs were not written for autonomous AI agents. A former SEC Honors Program intern in the SEC's Cyber Unit provides a ready-to-redline clause library organized by the KYA Five Pillars, with the regulatory basis and plain-language rationale for each provision.
You Cannot Delete What You Do Not Know the Agent Stored: The Memory Tool's Privacy Problem
Anthropic's Memory tool lets the agent decide what to store. A former SEC Honors Program intern in the SEC's Cyber Unit explains why that creates a novel data-mapping problem under CCPA and GDPR—and provides the remediation architecture deployers need before the first deletion request arrives.
What Claude Managed Agents Means for Your Compliance Stack: A KYA Framework Analysis
Anthropic's Managed Agents runtime shipped on April 8. A former SEC Honors Program intern in the SEC's Cyber Unit maps each product feature to the KYA Five Pillars governance framework—and identifies the compliance gaps deployers must close before their first production session.
Agents with Wallets, Agents in Vaults: A KYA Analysis of Claude Managed Agents for Digital Assets and Regulated Industries
Claude Managed Agents' architecture—persistent sessions, MCP connectors, bash execution, cloud-hosted containers—creates regulatory triggers self-hosted agents don't face. A former SEC Honors Program intern in the SEC's Cyber Unit maps the compliance gaps with the KYA Five Pillars framework.
AI Agent Liability in DeFi: Who's Responsible When the Bot Trades?
AI agents are executing autonomous financial transactions in DeFi right now. A former SEC Honors Program intern analyzes the products liability, securities, commodities, and AML exposure deployers face—with case law, enforcement precedents, and a compliance roadmap.
Deploying AI Agents or Building an AI Product?
Agent liability, governance frameworks, vendor contracts, and the state and EU rules that already apply — mapped to what you are actually shipping.
Talk to an Attorney