Carpenters’ Regional Council v. Wealthridge Construction (Ontario 2026): Labour Board Flags AI Use in Responding Party’s Submissions | Advocate Prakhar

Case at a Glance

Background

The Carpenters’ Regional Council, United Brotherhood of Carpenters and Joiners of America applied for certification to represent a craft bargaining unit in ICI and other sectors in Board Area 19. The responding parties — Wealthridge Construction Ltd., V&M Renovations Inc., and Timberline Mining Services Inc. — filed responses and subsequent submissions that the Board characterized as containing extensive allegations about the union’s conduct and improper motivations.

At a Case Management Hearing, the union raised the issue of artificial intelligence use in the preparing of the employer’s submissions. Vice-Chair Keating addressed the AI issue in his decision while also ruling on the preliminary status disputes and finding that the employer’s pleadings did not meet the Board’s established standards for establishing that the disputed workers were independent contractors rather than employees.

The Board held that the employer’s submissions about union conduct were irrelevant to the legal questions before it (no unfair labour practice complaint had been filed), and granted the union certification application.

The AI Issue

While the Board’s treatment of the AI issue was relatively brief compared to court sanctions decisions, the significance lies in the context: this is a labour tribunal — not a court — formally acknowledging and addressing the use of AI in party submissions. The Board noted the AI issue “briefly” as part of its procedural analysis, suggesting that labour relations boards and other administrative tribunals are beginning to develop their own frameworks for addressing AI-generated submissions. The duty of accuracy in tribunal proceedings is the same as in courts, but the procedural mechanisms for enforcing it differ.

What the Board Decided

• AI use in labour tribunal proceedings is now a formally documented issue — the Ontario Labour Relations Board treated the union’s flagging of AI use as a legitimate procedural concern requiring express acknowledgment.
• The employer’s submissions, characterized as “replete with allegations” about union conduct and motivations, were treated as irrelevant to the legal issues before the Board regardless of whether they were AI-generated.
• The employer’s failure to meet the Board’s established pleading standards for independent contractor status was determined on the existing record without a hearing, in part because of the inadequacy of the employer’s submissions.
• Certification was granted to the union on the basis that the disputed individuals were employees, not independent contractors.

“The Carpenters took issue with the responding parties’ use of artificial intelligence in the preparation of their submissions; I will address those issues, briefly.”

— Vice-Chair Neil Keating, Ontario Labour Relations Board, May 1, 2026

The India Angle

Indian Law Equivalent

India’s labour tribunals — including the Central Government Industrial Tribunal, National Industrial Tribunal, and Labour Courts under the Industrial Disputes Act, 1947 — operate under procedural rules that require accuracy in written submissions and pleadings. Section 11A of the Industrial Disputes Act empowers tribunals to manage proceedings and exclude irrelevant or improperly presented material. As AI use spreads in Indian labour practice, the OLRB’s approach of formally acknowledging and addressing AI use in submissions provides a useful template for Indian labour tribunals developing their own AI governance practices.

Bar Council Rules

Advocates appearing before labour tribunals in India are bound by the BCI Rules just as they are before courts. Rule 9 (not misleading the tribunal) and Rule 22 (factual accuracy) apply equally to written submissions before administrative and quasi-judicial bodies as before courts. AI-generated submissions that contain fabricated precedents, non-existent labour law provisions, or unsupported characterizations of the opposing party’s conduct would violate these rules in tribunal proceedings.

Practical Advice for Indian Advocates

• When using AI to draft submissions for labour tribunals, verify all cited decisions and statutory provisions with the same rigor applied to court filings — labour tribunal precedents are often not well-represented in AI training data, making hallucinations more common.
• If the AI generates allegations about the opposing party’s conduct or motivations, evaluate whether those allegations are actually supported by evidence in the record before including them — AI tools tend to generate plausible but unsubstantiated characterizations.
• Be aware that opposing counsel in labour proceedings is increasingly reviewing AI-generated submissions critically and may raise the issue before the tribunal — having a clean citation record is the best protection against such challenges.

Quick Takeaways

• Ontario’s labour tribunal formally acknowledged AI use in employer submissions in 2026 — a first for Canadian administrative proceedings.
• AI-generated submissions filled with irrelevant allegations rather than legal substance are treated as inadequate pleadings regardless of their source.
• The duty of accuracy in tribunal proceedings is the same as in courts — AI hallucinations in administrative filings carry the same risks.

Deep Dive: AI Governance in Labour and Administrative Tribunals

The Carpenters’ case extends the AI hallucination discourse beyond courts into administrative and regulatory proceedings. This extension matters because the volume of submissions before labour tribunals, human rights commissions, environmental boards, and other quasi-judicial bodies is vast — and those bodies are beginning to grapple with the same AI quality-control challenges as courts, often with less developed procedural infrastructure for addressing them.

One distinctive feature of the labour tribunal context is the party type. In Carpenters, the AI submissions came from an employer — not a pro se individual or a small law firm — in the context of resisting union certification. This suggests that AI use in litigation is not confined to resource-constrained litigants; even parties with access to legal advice are using AI-generated submissions in regulatory proceedings. The implications for the quality of administrative decision-making are significant: if the submissions before a tribunal are based on AI-generated mischaracterizations of the facts or law, the decisions issued in response may be correspondingly distorted.

The OLRB’s approach of “briefly” addressing the AI issue rather than imposing formal sanctions reflects the different procedural posture of labour tribunals compared to courts. Labour boards typically have inquisitorial elements, broad case management powers, and a focus on facilitating outcomes rather than punishing procedural misconduct. The Board’s approach — note the issue, assess whether it affected the outcome, and move on — may be appropriate for a first-generation response to AI in tribunal submissions.

For Indian practitioners advising clients before regulatory bodies — SEBI, CCI, NCLAT, TDSAT, and others — the Carpenters case is a prompt to review AI use policies for tribunal submissions. Indian regulatory bodies are among the most active adjudicatory forums in the country, and the scale of AI use in submissions before them is likely to grow rapidly. Developing AI submission policies now — before regulators begin imposing sanctions — is the prudent approach.