Attorney-Client Privilege and Cloud Software: What Every Attorney Must Know
- ABA Model Rule 1.6(c) — Duty to prevent inadvertent disclosure
- ABA Formal Opinion 477R (2017) — Cloud computing and confidentiality
- ABA Formal Opinion 498 (2021) — Virtual practice and security obligations
- ABA Formal Opinion 512 (2023) — Generative AI and confidentiality
When you drag a client document into Dropbox, upload a deposition to Otter.ai, or paste case facts into ChatGPT, you are transmitting potentially privileged information to a third-party server. The question every attorney should be asking — and most aren't — is whether privilege survives that transmission, and what obligations attach when it does.
The Core Duty: ABA Model Rule 1.6(c)
Rule 1.6(c) requires attorneys to "make reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client." The Comment to Rule 1.6 explicitly includes electronic means: "[w]hen transmitting a communication that includes information relating to the representation of a client, the lawyer must take reasonable precautions to prevent the information from coming into the hands of unintended recipients."
This duty did not go away when attorneys started using cloud tools. It transferred. The question isn't whether you can use cloud software — you clearly can. The question is whether you've taken "reasonable precautions" for the specific tool and the sensitivity of the information involved.
ABA Opinion 477R: The Cloud Framework
ABA Formal Opinion 477R (2017) is the most comprehensive guidance on cloud software selection. It establishes a five-factor framework for evaluating any cloud service:
ABA Opinion 512: Generative AI Changes Everything
ABA Formal Opinion 512 (2023) addressed generative AI directly. It makes clear that using AI tools for client matters requires the same confidentiality analysis as any other cloud service. Critically, it warns that many AI providers train their models on user inputs — meaning your case facts, your client's identity, and your litigation strategy could become training data for a commercial AI model.
Opinion 512 does not prohibit AI use — but it requires attorneys to understand how each tool handles data before using it for client matters. "The lawyer must act competently to safeguard information relating to the representation."
Evaluating tools: the ShieldDrop standard
- Uses your data to train AI
- Logs document content on servers
- No clear data deletion policy
- Subprocessors include major cloud platforms
- Free tier with no DPA available
- Zero-retention or in-browser processing
- Verifiable privacy claims
- DPA available on request
- No AI training on your data
- Clear incident notification policy
ShieldDrop Legal Suite was built on a single architectural premise: the safest cloud tool is one that never receives your data. By processing all documents in the browser via WebAssembly, ShieldDrop eliminates the disclosure risk entirely. There is no server transmission, no storage, and no subprocessor with access to client files. This satisfies the most stringent reading of Rule 1.6(c) for file processing tasks. Read the technical white paper →