New Canada Immigration And Refugee Rules Effective September 7

Last Updated On 3 September 2026, 9:13 AM EDT (Toronto Time)

Canada’s Immigration and Refugee Board is introducing new rules governing the use of artificial intelligence in immigration and refugee proceedings effective September 7, 2026.

Under the new Practice Notice, signed by IRB Chairperson Manon Brassard on July 10, 2026, parties appearing before the Board will be prohibited from using AI to generate or materially change personal evidence such as Basis of Claim narratives, affidavits and witness statements.

Certain other uses of AI will trigger mandatory disclosure requirements, including AI that generates or materially changes content in a document submitted to the IRB, while AI-assisted translation or transcription must always be disclosed.

The Practice Notice applies across all four divisions of the IRB: the Refugee Protection Division, the Refugee Appeal Division, the Immigration Division and the Immigration Appeal Division.

That means the rules cover refugee protection claims, refugee appeals, immigration appeals, sponsorship appeals, residency obligation appeals, removal order appeals, admissibility hearings, detention reviews and other matters within IRB jurisdiction.

An important distinction: these are rules for proceedings before Canada’s independent immigration and refugee tribunal. They are not blanket AI rules governing every Canadian immigration application.

The Practice Notice does not automatically apply to Express Entry applications, study permit applications, work permit applications, visitor visa applications or citizenship applications submitted to Immigration, Refugees and Citizenship Canada.

Those are separate processes handled by IRCC, not the IRB.

Also read: All The New Canada Immigration Changes Coming In 2026

AI Cannot Be Used To Create Or Rewrite Personal Evidence

This is the most consequential rule in the new Practice Notice.

Starting September 7, parties must not use artificial intelligence to generate or materially change content that reports or reflects a person’s evidence.

The IRB specifically identifies Basis of Claim narratives, affidavits and witness statements as examples of material that must be based on a person’s own knowledge and experience.

The Practice Notice lists the following as prohibited uses of AI:

Using an AI tool to draft a Basis of Claim narrative, affidavit or witness statement. Using AI to rewrite a person’s story. Using AI to create or alter photographs, videos, screenshots or other corroborative evidence.

This rule carries particular weight in refugee proceedings. The Basis of Claim narrative is the foundational document in a refugee protection claim.

It is the claimant’s personal account of why they are seeking protection, their experiences, their fear of persecution, and the specific events that led them to Canada.

Board members may consider inconsistencies between a claimant’s Basis of Claim narrative, testimony and other evidence when assessing credibility.

If AI generated or materially rewrote the narrative instead of it being based on the claimant’s own knowledge and experience, that would violate the new Practice Notice and could trigger the consequences set out by the IRB.

The same prohibition applies to affidavits and witness statements in IRB proceedings. These materials must reflect what the person actually knows and experienced, and AI cannot be used to draft or materially rewrite that personal evidence.

They must reflect what the person actually knows and experienced, not what an AI tool produced based on a prompt.

Spellcheck And Grammar Tools Are Still Allowed

The Practice Notice does not prohibit all technology assistance. Minor assistive functions such as spellcheck, grammar correction and formatting assistance remain permitted, provided those tools do not generate the evidence or make material changes to it.

The distinction is between minor assistance correcting a spelling error, fixing grammar, adjusting formatting and AI-generated or materially rewritten content.

Running a refugee claimant’s handwritten notes through a spell-checker is not the same as feeding their story into a generative AI tool and submitting what it produces.

The IRB has not established a rigid word-count or percentage threshold for what constitutes a “material change.” The test is whether the AI generated the content or materially changed the wording or substance.

New AI Declaration Required Starting September 7

If AI is used to generate or materially change text in a document submitted to the IRB, the party must include a declaration in the document.

The declaration essentially confirms that AI was used to create or materially change text, that all AI-generated content has been reviewed and verified by a named individual, and that the authenticity of any cited case law or other legal authorities has been checked.

A declaration is not required for minor assistive functions such as spellcheck, grammar correction or formatting assistance, provided those tools do not generate or materially change the wording of the document.

Parties may also choose to provide additional information about which AI software was used, why it was used, and whether AI was used for the entire document or only for specific paragraphs.

This additional detail is optional; the Practice Notice says parties “may” include it, not that they must. But the core declaration itself is mandatory whenever AI generates or materially changes text.

AI Translation And Transcription Must Be Disclosed

Use of AI for translation or transcription must always be disclosed. This is an especially important practical rule given the multilingual nature of IRB proceedings.

The IRB division rules already require translated documents to be accompanied by a declaration from a human translator.

If artificial intelligence is used to assist with the translation, the human translator must disclose that AI assistance in their declaration.

AI translation is not itself prohibited. The core requirement is disclosure and continued compliance with the applicable translation rules, including the human translator declaration.

AI translation does not eliminate the need for a human translator to stand behind the accuracy of the document.

Legal Cases And Citations Must Be Verified

The IRB’s Practice Notice addresses a well-documented risk with generative AI: the creation of false, fabricated or inaccurate information, including fictitious legal citations.

Parties remain responsible and accountable for everything they present before the Board. Where AI is used, parties must ensure the AI-generated content is accurate, verifiable and trustworthy.

The IRB specifically requires systematic verification of legal principles and case citations using reliable sources.

The Practice Notice identifies reliable sources as legal databases, official government websites, commonly referenced commercial publishers and trusted public services such as CanLII.

An AI-generated summary is not itself an adequate source for verifying legal authority.

Parties Must Be Ready To Explain Their AI Use

The new rules go beyond disclosure. Parties should be prepared to answer questions about how artificial intelligence was used in their proceeding.

When requested, they must be able to establish the authenticity and accuracy of their documents.

This means the September 7 regime is not a box-ticking exercise. The person submitting material to the IRB remains accountable for the underlying information.

If a Board member has questions about whether a document was AI-generated or whether the information in it is authentic, the party needs to be able to respond.

Do Not Use AI To Bulk Up Submissions

The IRB has directed parties not to fill their submissions with AI-generated information that does not relate to the particular facts of their case.

The Board’s message is clear: longer submissions are not automatically stronger submissions.

The Practice Notice cites Messa c. Canada (Citoyenneté et Immigration), 2025 CF 1557, in connection with this concern.

Padding a submission with irrelevant AI-generated material does not strengthen a case and may invite scrutiny.

Privacy Risks With AI Tools

The IRB warns parties to exercise caution when using AI tools to ensure sensitive information is kept confidential. Some AI tools do not have sufficient security features to protect information.

This warning is particularly relevant in refugee matters, immigration proceedings, detention matters and cases involving personal or sensitive evidence.

Refugee claimants, for example, may be disclosing details about persecution, violence or personal circumstances that could put them or their family members at risk if that information were exposed through an insecure AI platform.

What Happens If You Do Not Follow The New AI Rules?

This is where the Practice Notice has real teeth. If a party does not comply, the IRB says it may take the following actions:

Refuse to accept a document for filing or otherwise decline to rely on it. Draw negative inferences about the credibility of the evidence.

Disclose information concerning counsel to the appropriate professional regulatory authority under the IRB’s Policy on Disclosing Information Regarding the Conduct of Authorized Representatives to Regulatory Bodies.

Restrict or prohibit counsel from appearing before the IRB. Take any other action considered necessary.

These are possible measures the IRB says it may take depending on the circumstances. Not every violation will automatically trigger every consequence.

But the range of potential consequences is broad, from having evidence disregarded to having a lawyer or consultant barred from appearing before the Board.

An important distinction: consequences affecting counsel or authorized representatives, such as referral to a professional regulator or restrictions on appearing before the IRB, are directed at counsel.

An ordinary self-represented refugee claimant would not be reported to a professional regulatory body, though the IRB could still refuse or decline to rely on a document and draw negative inferences about the credibility of the evidence.

The Rules Cover All Four IRB Divisions

The breadth of this Practice Notice matters. The new AI rules apply across all four divisions of the Immigration and Refugee Board:

The Refugee Protection Division hears refugee protection claims and determines whether a claimant is a Convention refugee or a person in need of protection.

The Refugee Appeal Division hears appeals of RPD decisions, both from claimants and from the Minister. The Immigration Division conducts admissibility hearings and detention reviews.

The Immigration Appeal Division hears immigration appeals, including sponsorship appeals, certain removal order appeals and residency obligation appeals.

Any proceeding before any of these four divisions is subject to the new AI rules starting September 7.

What Changes On September 7?

Before September 7, the Practice Notice had been issued but had not yet come into force.

Parties already had existing legal and procedural obligations not to fabricate evidence or submit false information. Those obligations obviously continue. September 7 marks the commencement of new AI-specific rules.

Starting September 7:

The prohibition on using AI to generate or materially change personal evidence, including Basis of Claim narratives, affidavits, witness statements and corroborative evidence, becomes applicable.

The mandatory AI declaration requirement applies whenever AI generates or materially changes text in a document submitted to the IRB.

AI-assisted translation and transcription disclosure requirements apply. The obligation to systematically verify legal principles and case citations using reliable sources applies under the Practice Notice.

The potential consequences for non-compliance, including refusal of documents, negative credibility inferences and possible action against counsel, apply.

Who Should Pay The Most Attention

Anyone involved in an IRB proceeding should understand these rules, but certain groups need to pay especially close attention:

Refugee claimants preparing or revising their Basis of Claim narrative. People appealing refugee decisions to the Refugee Appeal Division.
Immigration appellants, including those appealing sponsorship refusals, removal orders or residency obligation findings.
People involved in admissibility proceedings or undergoing detention reviews.
Immigration lawyers representing clients before any IRB division.
Immigration consultants appearing before the IRB as authorized representatives. Other authorized representatives.
Self-represented parties preparing and submitting their own documents.
Translators preparing translated documents for IRB proceedings, who must disclose any AI assistance used in the translation.

The core practical takeaway is understanding the distinction between legitimate assistive AI use such as spellcheck and grammar correction and the prohibited generation or material alteration of personal evidence.

Also read: 10 New Canada Immigration Laws And Changes Coming In 2026

Frequently Asked Questions (FAQs)

What are the new Canada AI rules for immigration and refugee cases?

The Immigration and Refugee Board of Canada has issued a new Practice Notice governing the use of artificial intelligence in IRB proceedings, effective September 7, 2026. The rules prohibit using AI to generate or materially change personal evidence such as Basis of Claim narratives, affidavits and witness statements. AI that generates or materially changes content in a document submitted to the IRB must be declared, while AI-assisted translation or transcription must always be disclosed. Minor spellcheck, grammar correction and formatting assistance do not require a declaration when they do not generate or materially change the wording.

Can I still use AI for my immigration case after September 7?

AI is not entirely banned. Spellcheck, grammar correction and formatting assistance remain allowed, provided those tools do not generate evidence or materially change the wording or substance of a document. However, using AI to draft a Basis of Claim narrative, rewrite a person’s story, create or alter photographs, or generate other personal evidence is prohibited. Any AI use that generates or materially changes text in a document submitted to the IRB requires a declaration.

Do the new IRB AI rules apply to Express Entry or visa applications?

No, the Practice Notice governs proceedings before the IRB — Canada’s independent immigration and refugee tribunal. It does not automatically apply to Express Entry applications, study permit applications, work permit applications, visitor visa applications or citizenship applications submitted to IRCC.

What is the AI declaration requirement for IRB proceedings?

If AI is used to generate or materially change text in a document submitted to the IRB, the party must include a declaration confirming that AI was used, that the content has been reviewed and verified by a named individual, and that cited case law and legal authorities have been checked for authenticity. A declaration is not required for minor assistive functions such as spellcheck or grammar correction.

What happens if I use AI without disclosing it to the IRB?

The IRB may refuse to accept the document for filing, decline to rely on it, draw negative inferences about the credibility of the evidence, report counsel to their professional regulatory authority, restrict or prohibit counsel from appearing before the Board, or take other necessary action. These are potential consequences the IRB may apply depending on the circumstances.

Does AI translation still require a human translator declaration at the IRB?

Yes, the IRB division rules require translated documents to be accompanied by a declaration from a human translator. If AI assisted with the translation, the human translator must disclose that assistance. AI translation does not eliminate the requirement for a human translator declaration.

Fact-Checked: This article is based on the official Practice Notice: Use of artificial intelligence in IRB proceedings, published by the Immigration and Refugee Board of Canada, signed by Chairperson Manon Brassard on July 10, 2026, and coming into force on September 7, 2026. All substantive claims were verified against this primary source as of September 3, 2026. Additional sources include the IRB’s procedures and practice notices page, the Policy on Disclosing Information Regarding the Conduct of Authorized Representatives to Regulatory Bodies, and official IRB division pages.

Disclaimer: This article is for informational purposes only and does not constitute legal advice. Readers involved in IRB proceedings should consult a qualified immigration lawyer or authorized representative for guidance specific to their case. Anyone appearing before or submitting material to the IRB from September 7 onward should understand what AI use is prohibited, what must be disclosed, and that responsibility for the accuracy and authenticity of material remains with the party.

Kamal Deep Singh, RCIC

Kamal Deep Singh, RCIC (Regulated Canadian Immigration Consultant) licensed by CICC (formerly known as ICCRC) with member number R708618. He brings extensive knowledge of immigration law and new changes to rapidly evolving IRCC.