Picture a lawyer working against a filing deadline. A client contract needs a fast summary, so pasting it into a cloud AI chatbot looks like the obvious shortcut — fast, convenient, and the output would be genuinely useful. It would also be a decision with confidentiality implications worth thinking through before it happens, not after.
The question isn't whether AI is useful for legal work — it clearly is. It's where the processing happens.
What Happens When You Upload a Client Document
Cloud AI tools require your file to leave your device and travel to a third-party server for processing. Once it does, you're relying on that provider's infrastructure, retention policy, and security practices — not your firm's. See our full breakdown of what happens to uploaded documents for the mechanics.
The Privilege Question
Attorney-client privilege can be waived by disclosure to a third party. Whether routing a privileged document through a cloud AI vendor's servers constitutes that kind of disclosure isn't settled uniformly — it depends on jurisdiction, the vendor's terms, and the specific facts. This is genuinely unresolved territory in several jurisdictions, which is exactly why the safer default is to avoid the question altogether by keeping the document off third-party servers in the first place.
The practical takeaway: the cleanest way to avoid privilege and confidentiality questions is to not create them — keep the document off any server you don't control.
What On-Device AI Changes
An on-device AI tool processes documents locally: parsing, embedding, retrieval, and generation all happen on the Mac, iPad, or iPhone in front of you, with no client file transmitted anywhere. This doesn't just reduce risk — it removes an entire category of "what does the vendor do with this" questions, because there's no vendor in the data path.
Kynora AI's retrieval-augmented generation pipeline is built to run entirely on-device: import a case file in pdf format, ask a plain-language question, and get an answer with a citation to the source section and page — so you're verifying against the original document, not trusting a paraphrase.
Where This Matters in Practice
- Reviewing a long agreement for where a specific term appears
- Finding a specific topic in a long document
- Working on documents without reliable internet or no internet
- Making scanned productions and legacy PDFs searchable via on-device OCR
None of this replaces professional judgment, and none of it is legal advice — always verify any AI-assisted answer against the source document, and check with your firm's ethics counsel before adopting any AI tool for client work.
Kynora AI keeps documents like these on-device from import to answer — no client file transmitted anywhere.