From the Public Safety edition of October 21, 2026
Canadian Judicial Council’s second-edition AI guidelines list uses judges must not delegate to AI, bar court documents in non-court tools, and treat AI transcripts as personal notes
Canadian Judicial Council · Courts / judicial AI guidance · Canada
- Publisher
- Canadian Judicial Council
- Original publication
- October 20, 2026 (Canadian Judicial Council)
- Source retrieved
- 2026-10-22
- Event date
- 2026-10-20
What happened
The second edition replaces the 2024 guidelines and sorts judicial AI uses into permissible, high-risk borderline and inappropriate. Inappropriate uses include entering court documents into non-court tools (naming ChatGPT, Claude, Grok, NotebookLM and others), assessing credibility, resolving disputed facts or law, exercising discretion, and deciding guilt, liability or sentence. Borderline uses include identifying legal tests and sentence ranges. AI transcripts of proceedings should be treated as personal notes, not official records. Individual judicial use does not ordinarily require disclosure, but courts should explain how deployed systems are evaluated and governed. Court administration AI requires judicial oversight even on government platforms.
Why it matters
One of the most specific judicial AI use taxonomies published to date, naming consumer tools and separating individual judge use from court administration systems. U.S. state courts drafting AI policies can compare its borderline category, such as sentence ranges, with their own rules.
Evidence and measured results
Opened the CJC news release (Oct. 20) and the full guidelines PDF; read the transparency, court administration, sensitive information, permissible, borderline and inappropriate use sections and transcription guidance.
Limitations and uncertainty
Guidance, not binding rules; examples may migrate between categories as models change (per the document); no compliance measurement.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-22; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Courts need approved AI tools and clear rules as judges face pressure to use consumer AI.
- Stakeholders
- Chief justices
- Court administrators
- Court IT
- Judicial education bodies
- Discovery questions
- Which AI tools are judges using today?
- Is there a court-approved environment?
- How are AI transcripts kept apart from the record?
- Who approves court administration AI?
- Value hypothesis
- Mapping uses to clear categories lets a court approve low-risk assistance while blocking prohibited uses.
- Potential engagement
- AI use classification and policy alignment review for one court system.
- Unsupported claims
- That the guidelines are binding
- That compliance reduces hallucinations
- That any product is endorsed
Pre-sales engineering
Role takeaway
- Fit
- Courts deploying AI for summarisation, research or transcription.
- Architecture and integration
- Court-controlled AI environment separated from shared government platforms where independence requires.
- Prerequisites
- Data classification of court records
- Approved tool list
- Logging
- Constraints
- Canadian context
- Shared government platforms
- Sealed and publication-ban data
- Security
- Prevent training on judicial data and keep sensitive categories out of external services.
- Proof of value
- Proposed: verify no court data leaves approved environment
- Proposed: confirm AI transcripts are labelled non-official
Delivery
Role takeaway
- Work
- Classify AI uses, set controls per category, train judges and staff, and publish system descriptions.
- Dependencies
- Judicial sign-off
- IT inventory
- Training capacity
- Ownership
- Chief justice office with court administration
- Skills and adoption
- Judicial training on verification and confidentiality.
- Governance checkpoints
- Judicial approval for each court administration AI system
- Periodic review of category examples
- Acceptance criteria
- Proposed: all deployed AI systems publicly described
- Proposed: training completed by judges and staff
- Risks
- Shadow use of consumer tools
- Category drift as models change
- Vendor training on court data
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
Court-approved AI systems should be used for non-public information; government-wide shared platforms do not by themselves resolve independence or confidentiality questions.
Governance
Who approves, reviews and stays accountable for outcomes?
Judicial leadership over procurement, configuration and deployment of court administration AI; institutional transparency about deployed systems.
Security and privacy
What data, permissions and controls need testing?
Lists sensitive categories (draft reasons, sealed records, privileged material, publication bans) to keep out of AI systems; concerns about training on judicial information.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Calls for training of judges and court staff on prompting, verification, confidentiality and security.
Procurement
What should contracts, pricing and exit terms secure?
Contractual confidentiality assurances do not relieve judges of responsibility; data sovereignty should inform tool selection.
Operating model
Which teams own the service once it runs?
Separates individual judicial use from institutional court administration systems with different governance.
Publication history
- 2026-10-21Public Safety · Issue 438 resources
Stable resource ID: canadian-judicial-council-ai-guidelines-second-edition-impermissible-uses-20261020