Education Plan Cites Nonexistent Sources

On August 28, 2025, Education Accord NL released its 10-year education roadmap for the Canadian province of Newfoundland and Labrador. The 418-page document took 18 months to complete and came with 110 different calls to action.

One of those calls specifically addressed generative AI. It encouraged the government to “provide learners and educators with essential AI knowledge, including ethics, data privacy, and responsible technology use.”

However, shortly after the report’s release, educators noticed discrepancies in the report’s citations. Specifically, they found that at least 15 of the sources cited in the report are fabricated.

This has raised significant questions about the report, specifically whether AI was used to generate some or all of the report.

In a statement, the Department of Education and Early Childhood Development, which commissioned the report, acknowledged that there was a “small number of potential errors in citations” in the report.  It said it is addressing the errors and that the report will be updated afterward.

However, neither the Department of Education and Early Childhood Development nor Education Accord NL has answered questions about whether AI was used in the report or how the false citations got in.

As a result, the report is facing stiff criticism, with many noting the irony of (potentially) using AI in an unethical way in a report largely about responsible technology use.

But should the report be scrapped? That’s a difficult question, especially when we don’t know exactly what happened.

What We Know

The report was unveiled on August 28 by Anne Burke and Karen Goodnough, the co-chairs of the Education Accord NL, and Education Minister Bernard Davis.

The report was commissioned by the Department of Education and Early Childhood Development, which sought a plan to help modernize its education system.

However, Assistant Professor Aaron Tucker began to review the report and noticed that some of the citations were fabricated. Tucker, along with Burke and Goodnough, is an instructor at Memorial University. According to Ars Technica, Tucker’s research focuses on AI history in Canada and has taken a strong interest in these fabricated sources.

One of the worst offenders is a citation of a 2008 National Film Board movie entitled Schoolyard Games. The movie does not exist, but it was used on the University of Victoria’s style guide as a fake example to teach citation formatting.

According to Education Minister Bernard Davis, the citations do not impact the report’s veracity. He notes that only 15 of the 450 sources have been identified as false, which represents roughly 3%.

“I don’t think it impacts the credibility of the document,” he said in an interview with the CBC. Burke and Goodnough released a joint statement noting the “errors” and saying that they were updating the report. The statement went on to say that:

“It was co-authored by us as co-chairs, with significant written input from pillar co-chairs, pillar committee members, advisory committees, and the thousands of voices from throughout Newfoundland and Labrador who shared their input and experiences. Any suggestion otherwise is inaccurate.”

However, it’s not that simple. Critics of the report have some valid questions that remain unanswered.

Sadly, it doesn’t look like we’re going to get those answers anytime soon.

The Big Problem

On one hand, it would be easy to follow Davis’ approach and highlight that only 3% of the citations are at issue. However, that would require ignoring three very serious problems.

First, putting fake sources into a work is fabrication. Fabrication is a very serious research integrity sin and is often considered more problematic than simple plagiarism. Although one can attribute plagiarism to sloppiness or inattentiveness, fabrication is almost always a deliberate and intentional deception.

If this were published in an academic journal, it would almost certainly be subject to retraction. Fabrication is not tolerated in research, and the only difference is that this is a government report, not a study in an academic journal.

The second issue is that we don’t know what other issues the paper may have. We know of 15 sources that were fabricated. However, that doesn’t mean the other is not valid. Just because a source is real doesn’t mean it says what the report claims it does.

A thorough analysis of the paper should be conducted by a third party. It needs to verify that the rest of the citations are accurate, that the wording is original and that no other authorship or citation issues are present.

Finally, and most importantly, there’s the issue of AI. As many others have noted, fabricated sources are a hallmark of AI-generated work. I am inclined to agree. AI is the most likely explanation for how those sources got in there. The public and the government have a right and need to know when, if and how AI was used. Right now, no AI use has been disclosed.

But there’s no good answer here. Either A) They used AI in a sloppy manner and failed to disclose it, or B) the authors put the fabricated sources in the report themselves. Neither answer is good, and both cast serious doubts on the reliability of this study.

Sadly, it appears that those involved want to handle the matter internally and without transparency. They’ve already stated that they will update the report with corrections, and, most likely, that will be the end of it.

The lack of accountability and transparency is worrisome. Even if you agree that these issues can be properly addressed with corrections, we first need to ensure that these are the only issues.

Without that, any rush to vindicate the report is premature.

Bottom Line

In the end, I have to agree with the report’s critics and say that this undermines the report. However, to my mind, the errors aren’t as worrisome as the response to the allegations. While I understand the desire to circle the proverbial wagons, this is a time when transparency and due diligence should take the lead.

As bad as some fabricated sources are, it’s not nearly as serious as sweeping the issue aside. One challenges the authenticity of a single study; the other breeds mistrust in academia as a whole.

Simply put, this high-level report is getting less integrity oversight than what many of the province’s students endure on a daily basis. If a student had fabricated some of their sources on a paper, possibly using an AI without permission or disclosure, there would be repercussions.

The government owes its constituents transparency. Conducting a thorough investigation of the work and ensuring that it meets the standards of academic integrity is a tedious but crucial part of the process.

After all, academia relies on public trust. A prominent report with fabricated sources is bad enough in that regard, but not following up on it compounds the issue.