It’ll be great when they work out the quirks in AI


For instance, the Arabian Journal of Geosciences has a problem: their articles are too obviously fake.

Some titles of the farkakte research: “Simulation of sea surface temperature based on non-sampling error and psychological intervention of music education”; “Distribution of earthquake activity in mountain area based on embedded system and physical fitness detection of basketball”; “The stability of rainfall conditions based on sensor networks and the effect of psychological intervention for patients with urban anxiety disorder.” A complete list of the retracted papers can be found here.

They read a bit like a college student throwing around big words to cover up a lack of understanding. Though purportedly written by humans, the content of each paper definitely reads as if it were put together by a computer that doesn’t quite grasp speech patterns or grammar. The papers are filled with redundancies and generally lack logic.

Right away, I noticed a problem: they should have used the more formal German “verkakte” rather than the alternative Yiddish spelling. Oh, right, and the paper titles are absurd, too.

I see two sources of problems: institutions that demand frequent publications, even where it isn’t warranted, and extremely lazy journal editors who rubberstamp everything.

As amusing (or alarming) as the idea of earthquakes being connected to basketball might be, the screwup highlights issues in science publishing that let farcical research slip into the realm of real work. As highlighted by the Chronicle of Higher Education in August, when the 400-odd papers in the geosciences journal got expressions of concern attached to them, many suspicious papers appear to have been written by scholars affiliated with Chinese institutions, where researchers are incentivized (sometimes financially) to publish in notable journals and where many doctoral students must publish a paper before graduation. The founder and editor-in-chief of the Arabian Journal of Geosciences told the Chronicle at the time that he reads every paper published in the journal each month (which would mean about 10 papers per day, including weekends), and that he thinks the fabricated research got into the journal through hacking.

Sure. Abdullah M. Al-Amri, editor-in-chief of the journal, reads every submission, and all those ridiculous papers must have been hacked into place. He reads every article, except he never takes a look at the journal once it’s been published. And he never reads the correspondence from legit researchers who point out the kind of crap getting splattered all over the pages. He just failed to notice that “Structure of plain granular rock mass based on motion sensor and movement evaluation of dancers” got published.

Just wait until Chinese researchers discover ChatGPT. We desperately need our pseudoscientific garbage to be more readable.

Comments

  1. says

    Just wait until the chatAI manages to produce original, useful research, purely by accident.
    Autogenerating data might be a little more tricky.

  2. René says

    they should have used the more formal German “verkakte” rather than the alternative Yiddish spelling.

    Both ‘verkakte’ and ‘farkakte’ seem to be Yiddish. Deutsche Rechtschreibung would have it as verkackte.

  3. Tethys says

    I read verkackte* as OHG, which is the source for Yiddish. It’s definitely a word I’ve heard used, generally accompanied by whatever the verkackte item might be getting chucked into a trash bin along with a string of growling German curse words like sheiss, and svine.

    to be shat upon

  4. lasius says

    No, Tethys. Just like whatever you claim to speak isn’t based on OHG, neither is Yiddish. Yiddish developed out of MHG.

    Also the words you are looking for are “Scheiß” and “Schwein”.

  5. Tethys says

    @Lasius
    Yiddish developed from WesternGermanic/High German starting at least 1600 years ago. (Possibly much earlier, but there are very few records that document Jewish people in Europe) There was a synagogue in Cologne circa 400, well before there is any such thing as MHG. That is a time period, not an actual dialect of German.

    Danké for the proper spelling of common German expletives. This device lacks a few of your letters entirely, and I freely admit my ability to spell properly in German is scheiss.

  6. lasius says

    @Tethys.

    Yes, a synagogue, maybe. But 1600 years ago Latin was spoken in Cologne and Yiddish developed further south in the region of Worms and Speyer anyway. While there were Jews in Germany much earlier, Yiddish didn’t split off from German untill the high middle ages and the time of MHG. This is easily demonstrated by the shared sound changes that Yiddish and MHG have and that Yiddish would not possess if it already split off in the time of OHG.

  7. simplicio says

    I asked ChatGPT how the structure of plain granular rock mass based on motion sensor and movement evaluation of dancers works. Guess what. Mr. Abdullah M. Al-Amri needs to turn his editorial position over to ChatGPT.

  8. Tethys says

    This is easily demonstrated by the shared sound changes that Yiddish and MHG have and that Yiddish would not possess if it already split off in the time of OHG.

    I really don’t care about it’s development, it STILL originated in OHG, schmuck. Otherwise I wouldn’t be able to understand a verkackte thing about Yiddish, yet I’m quite comfortable with schlemph, and mensch, and multiple other German words that have been incorporated into Yiddish. Conversely it’s easy enough to read if you know OHG to modern German, because that’s usually how it’s spelled. (But not pronounced)

    So many throat gravel aka gutturals!!! Ich, challah, schlemiel, shule, leibchen.

  9. lasius says

    I read verkackte* as OHG

    Your earlier quote makes no sense then, since “Kacke” and “verkackte” are not attested in OHG. I am sure the words ancestor must have existed but we don’t know what exactly it was since early medieval monks apparently didn’t care to use this word within their texts.

    But in any case, saying that Yiddish originated in OHG, when it clearly originated in MHG is, while technically true from a certain angle, a weird thing to say.

    So many throat gravel aka gutturals!!! Ich, challah, schlemiel, shule, leibchen.

    German is no more guttural than English, you just percieve it that way. “Challah” and “Schlemiel” are not a words derived from German but Hebrew. And I don’t think “Leibchen” (a type of shirt) is a word in Hebrew. The only word of your list pronounced whith a sound that could be decribed as “guttural” is “Challah”, the others only use sounds produced in the front or center of the mouth.

    Conversely it’s easy enough to read if you know OHG to modern German

    You don’t know OHG though. OHG is ancient and attested to us in only comparatively few monastic sources. If you really know OHG then you are a time traveller and I implore you to contact some linguists to have your knowledge of OHG recorded.

    And don’t kid yourself, you don’t know modern German either as I can clearly see from your constant mispellings,which are telling me you don’t know the correct pronunciation of these words either.