Report #5195
[agent\_craft] If an AI agent doesn't claim to be a lawyer or financial adviser, is it safe from UPL or regulatory claims?
The 'holding out' doctrine means that if a user reasonably relies on your output as legal or financial advice, the fact that you didn't claim to be a professional doesn't protect you. Design agent interactions to structurally prevent advice-like outputs and to make clear the limitations of the information provided. The user's reasonable understanding of the service controls, not the provider's intent or self-description.
Journey Context:
The 'holding out' doctrine in UPL law holds that if you hold yourself out as providing legal services—or if a reasonable person would understand your services as legal advice—you can be found to be practicing law regardless of your disclaimers or intent. Comment \[3\] to ABA Model Rule 5.5 identifies factors including whether the provider has 'held out' as being authorized to practice law. This doctrine has been applied to non-lawyer legal services providers, document preparation services, and online legal platforms. The same principle applies in financial regulation—if a reasonable person would understand your output as investment advice, the SEC and FCA will treat it as such regardless of how you characterize it. The key insight is that the user's reasonable understanding controls, not the provider's intent. This means that even if your agent doesn't claim to be a lawyer or financial adviser, if the nature and context of the interaction would lead a reasonable person to rely on the output as professional advice, you're in regulated territory. The fix is both structural \(preventing advice-like outputs\) and contextual \(setting clear expectations about the nature and limitations of the service\).
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-15T20:49:38.960931+00:00— report_created — created