LatestAI & the law7 min read
How AI is actually changing legal work
Under the noise, one shift matters: machines that can read a private corpus and answer with sources. Here is what changes, what does not, and where it lands first.
Read the postVaultr Journal
Notes from the boundary
Engineering notes, verification methodology and honest compliance writing from the people building Vaultr. No newsroom theatre — every post traces to work that exists in the repository or on this site.
12 posts · updated 15 September 2026
Filed underAI & the law · 3Practice · 2Engineering · 3Briefings · 2Compliance · 1Verification · 1
The research desk
Sources we read, so you can read them too — not summaries of them.
Primary documents only: the acts, frameworks, licences and reports behind most legal-AI commentary. Descriptions below are ours; the documents belong to their publishers.
- AI Index ReportStanford Institute for Human-Centered AIThe neutral annual baseline for AI capability and adoption claims — useful for discounting vendor superlatives, including ours.
- EU AI Act — Regulation (EU) 2024/1689EUR-LexThe legal text itself. Obligations for AI systems, transparency provisions and timelines, straight from the official journal.
- AI Risk Management FrameworkNISTA voluntary structure for governing, mapping, measuring and managing AI risk. Maps well onto a procurement checklist.
- OpenChain conformanceOpenChain ProjectThe official description of the conformance paths — including self-certification — that our trust-center statement relies on.
- GNU Affero General Public License v3.0GNU ProjectThe licence text governing Vaultr's public repository. Read the licence, not summaries of it.
- Personal Data Protection CommissionGovernment of SingaporeThe regulator for Singapore's PDPA — our home jurisdiction — including advisory guidance relevant to AI and data use.
Follow the journal
New posts land here first. The feed is a plain RSS file — no subscriber list, no tracking pixels, no email required.