Canada’s GCtranslate AI Tool Plagued by Errors in First Six Months

FOI Documents Expose Flaws in Federal AI Translator

The Canadian federal government’s newly deployed AI translation tool, GCtranslate, is reportedly struggling with accuracy just six months into its operation. According to documents obtained by CBC/Radio-Canada through a Freedom of Information Act request, the system has generated a concerning volume of errors in official government documents.

These mistakes range from minor linguistic mishaps to significant alterations that could compromise the integrity of public records. The reporting highlights specific instances where the AI failed to maintain context, particularly when translating from English to French.

Risks to Official Communications and Public Trust

For citizens relying on accurate bilingual information, these errors represent more than just amusing glitches; they signal potential failures in critical infrastructure. One documented instance involved the abbreviation “TB” for the Treasury Board being translated as “Tuberculosis,” while another saw a person named “Mike” changed to “Stephanie.”

Such hallucinations—where AI invents details not present in the source text—undermine the reliability of automated tools intended to streamline government operations. The phrase “Buy Canadian Policy” was mistranslated as “Achetez une politique canadienne” (Buy a Canadian policy), demonstrating how literal interpretations can distort meaning.

This situation underscores the gap between current large language model capabilities and the precision required for legal and administrative contexts. While AI offers speed, the cost of correcting downstream errors in official documentation may outweigh initial efficiency gains.

Unanswered Questions on Oversight and Correction

It remains unclear whether the government has implemented additional human oversight layers or if GCtranslate will be suspended pending further review. The incident serves as a cautionary tale for other organizations adopting generative AI for high-stakes tasks without rigorous validation protocols.

As federal departments continue to integrate these tools, the balance between automation and accountability becomes increasingly vital. Without transparent correction mechanisms, trust in digital public services risks erosion.