Google's agentic AI and legal liability: who bears the risk when the agent gets it wrong?
Google's new AI agents (I/O 2026) act without direct human supervision. What liability chain applies to the business that deploys them?
The announcement at Google I/O 2026 of agents capable of executing complex end-to-end tasks — booking, buying, contracting, arbitrating — moves the question of legal liability far beyond the simple use of a software tool. A business that delegates a legally binding decision to an agent exposes itself to contractual, tortious and potentially criminal liability, often through a chain of causation that is hard to read.
The first reflex is to hide behind
The first reflex is to hide behind the provider's terms of use. Yet these operate massive exclusions: no guarantee on outputs, derisory liability caps, and blanket referral to user responsibility for anything the agent initiates. In practice, the residual risk sits with the client.
The first reflex is to hide behind the provider's terms of use.
Three axes should structure governance
Three axes should structure governance: mapping the processes in which an agent acts without human validation, documenting the supervision logic (logs, thresholds, stop points), and adjusting professional liability insurance. Poor traceability becomes an aggravating factor in court.
In the medium term, the European framework
In the medium term, the European framework will likely converge toward strict operator liability, along the lines of the AI Liability Directive. Anticipating this shift means installing an agentic governance today that goes beyond legal, bringing in IT, compliance and business leaders.
Key takeaways
- 01A business that delegates a legally binding decision to an agent exposes itself to contractual, tortious and potentially criminal liability, often through a chain of causation that is hard to read.
- 02In practice, the residual risk sits with the client.
- 03Poor traceability becomes an aggravating factor in court.
- 04Anticipating this shift means installing an agentic governance today that goes beyond legal, bringing in IT, compliance and business leaders.
Published on
24 June 2026
Section
Law & AI
Rackham Limited
Take this further
A confidential conversation with the Rackham team to translate these questions into your organisation.
Start the conversation →Related articles
Continue reading →
Compliance
AI Overviews in Europe: compliance, IP and the risk of misinformation
As Google's AI summaries roll out in France, what obligations apply to businesses embedding these tools in their processes?
Data & AI
Gemini Omni and sensitive data governance: when multimodal AI crosses boundaries
Gemini's ability to process text, image, audio and video in a single flow raises questions of data classification and protection.
M&A
M&A and due diligence in the age of AI: how agentic search reshapes strategic audits
How autonomous research tools reshape the exploitation of target data and the detection of legal risks during diligence.