Steve Vainker

The irony:

"Overall, this study provides valuable insights into the..."

Google Scholar before 2021: 35
Since: 2,385

Sadly, this is on of the better articles on AI published in the journal. They published 16 on AI that were written by...AI.

bsky.app/profile/step...

As I say, I am highly confident each of these is substantially AI-written.

My hope is that a more focused and relevant BERJ will emerge from this.

doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...

doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...

doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...

And the rest:

doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...

<em>British Educational Research Journal</em> | BERA Journal | Wiley Online Library The increasing integration of artificial intelligence (AI) in education has led to a surge of interest in AI-assisted learning environments. These environments offer various advantages, yet a deeper ...

doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...

<em>British Educational Research Journal</em> | BERA Journal | Wiley Online Library The incorporation of artificial intelligence (AI) into education has drastically transformed language learning, particularly for English as a foreign language (EFL) students, by introducing innovativ...

doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...

The full evidence for each one is on my X (same name). In around 3-4 cases, there were fake references (in one case, 21).

Here's the full list of 40. First, 16 articles that were on AI:

doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...
doi.org/10.1002/berj...

My logic here is that if 1-5 of such phrases are used, it maybe doesn't explain much, but the odds of a human writing 15-20 such phrases become extremely small. Supported by further evidence (Winston) then I am highly confident of making the claim.

e.g. "effectiveness in capturing the nuances of"
Before 2023: 0
Since: 151

If I find 15-20 instances of phrases rarely or never used before 2022/3, which have now become common, AND Winston AI (which uses distinct methods) flags it as AI, then I regard it as highly likely as being AI.

In the first instance, I use Winston AI, an AI checker. If it flags it as human, I do not investigate any further (in one case it was flagged as mostly human, but there were fake references, so I included).

Next, I look for fake references and check for a preponderance of post-2022 phrases...

And more generally, the AI articles (and many others) do not advance the field in any significant way (at all). They are formulaic and over-obsessed with details at the expense of a main point.

So, in terms of the evidence for the AI articles, I have a two-stage process.

This one (£) gave contradictory information about ethical approval

doi.org/10.1002/berj...

Despite being an experiment about cross-cultural awareness, this paper (£) had no mention at all about its ethics.

doi.org/10.1002/berj...

While this paragraph reads like a Microsoft advertorial

doi.org/10.1002/berj... (£)

There are many more, but in this AI-written paper the same passage is repeated verbatim three times in a row.

doi.org/10.1002/berj...

Also, they're not informed on ed research...re teaching styles, they claim an 'overwhelmingly focus on higher education, with much less evidence on the use and effectiveness of LBT [i.e. passive] vs. DBT [i.e. active] in earlier education like during mandatory schooling'. This is an absurd comment!

<em>British Educational Research Journal</em> | BERA Journal | Wiley Online Library We investigate the effects of ‘lecture-based’ (LBT)—i.e. individual work and rote learning—vs. ‘discussion-based’ (DBT)—i.e. participative and focused on student-centred learning—teaching styles on t....

A further paper (£) was organised around framing teachers into two camps - 'lecture based teaching' and 'discussion-based teaching' based on three questions that they answered in a questionnaire - except the authors do not share the questionnaire with us.

doi.org/10.1002/berj...

This isn't even the worst thing about their lit review, which also extensively cited a masters thesis (Garcia, 2016) to support claims about peer support, but the thesis wasn't actually about peer support. Read this sorry tale from the thesis (I love the 'known unreasonable assumption' comment).

Extract from Bas & Xu, 2024

For over a year, amid some excellence, the journal has been publishing a significant amount of low-quality research. For example, see this from Bas & Xu, 2024 (£) 👇 (Ps, Xu does not claim this, and of course De Jong doesn't intend it as a universal claim [note date])

doi.org/10.1002/berj...

As the main journal, the British Educational Research Journal needs to be at the forefront of this. I'm not saying that every article needs to directly inform policy and practice in the UK, but it does need to showcase the deeper expertise that academics can offer.

For expertise related to educational research, the Government in England has looked away from universities. This may or may not be correct, but it means that the British Educational Research Association has a job to do to re-establish university academics as the foremost source of expertise.

I have discovered 40 (forty) articles published by the @berj-2025.bsky.social that look highly likely to have been written by AI. Before I explain my evidence & list the articles, I wanted to explain why it matters.

The BERJ is a publication I loved and would read through. It can become that again

Stephen Vainker on X: "🧵HUGE unsupported & often absurd claims: 'Teachers give students numerous homework assignments in classes almost every day (Cooper, 2007); however, students rarely do the homework assigned to them (Xu, 2011).' The Xu (2011) reference doesn't say this https://t.co/1zH2HQWOss https://t.co/Jw3fF57N8P" / X 🧵HUGE unsupported & often absurd claims: 'Teachers give students numerous homework assignments in classes almost every day (Cooper, 2007); however, students rarely do the homework assigned to them (Xu, 2011).' The Xu (2011) reference doesn't say this https://t.co/1zH2HQWOss https://t.co/Jw3fF57N8P

This is far from the first time very low quality research has appeared in the BERJ in the last 18 months. This is perhaps the worst literature review I've ever read:
x.com/StephenVaink...
I'll write a full review of a number of no-quality recent BERJ papers in the near future.

Zou, D., Xie, H., & Wang, F. (2021). Inclusiveness and fairness in AI-driven education: Challenges and strategies. AI in Education Review, 15(1), 34–52.

Zhao, Y., Wang, L., & Lee, H. (2022). The role of chatbots in language learning. Journal of Language and Technology, 19(3), 112–129.

Zhao, Y., Yin, H., & Wang, L. (2022). Ethical concerns in the application of AI in language education. International Journal of Educational Technology, 19(2), 89–104.

Loewen, S., Isbell, D., & Spino, L. (2020). The effectiveness of AI-driven tools in language learning: Addressing ethical and pedagogical issues. Computer Assisted Language Learning, 33(7), 845–863.

Lau, L., Hong, S., & Lin, D. (2022). Data privacy and algorithmic bias in language learning AI systems. Journal of Educational Technology, 17(4), 210–223.