I wanted to add another comment to note that the person I was talking back and forth with here appears to have deleted their account.
I would expect their comments to remain in place, but they are missing now.
So, it's looks like I'm having a bit of a crazy conversation only with myself, which wasn't the case.

I believe this is because a provider has given that record metadata to Ex Libris in English.
There is the record in English from ProQuest, and this matches to the ProQuest source platform showing the record metadata in English and the article in Chinese: cdi_proquest_journals_1858231693
There is also a record which has Chinese metadata also, rather than English. This is cdi_hyweb_hyread_00440336 and I believe this is from the Chinese Electronic Periodical Service (CEPS).
If you search by the Chinese characters, the Chinese version record will be returned in Primo.
Basically, Ex Libris is presenting what the provider has given them, which does seem reasonable and right to me and is their standard practice to not adjust provider metadata.
They also don’t Match & Merge when there are language variations, which does seem right to be also as that could get very messy trying to merge such varying data.
Are you thinking that Ex Libris should add a detection factor during provider feed ingestion to identify if the record metadata is in English, but there is Language metadata stating something else like Chinese, and try to employ a translation service to transform the data to that language?

We have access to this example content via a CDI record from ProQuest, same as yours. The metadata is in English, and it has a Language display field of Chinese.
When navigating to ProQuest, I see this is replicated by the English metadata, Chinese Language indication, and the pdf in Chinese.
In looking around your site, I see that you have a Language facet but in a random check of your records I can’t see any Language fields displayed in records. When using this facet to limit to Chinese after doing a search by the article title, this record appears, Have you deliberately disabled the Language display field? If not, I think it is a defect that you are not seeing the Language display field, to make clear the article is in Chinese on the source platform.

Is there an example of this?
Whenever I have seen this, it has appeared to be explainable by match and merge of multiple sources and the full text source content reflects the same. For example, the title in English in the source but the article in another language and in the Primo record the title is in English.
Or is this asking instead for something like that Ex Libris step in to change the metadata themselves for the title to match the article language despite what the provider may have made available to them, such as by AI generative tools?