AI influenced worry dreams
Jul. 31st, 2023 12:42 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
kareinaThe thesis chapter I am working on right now is a summary of the steatite items held in Swedish museums. A tedious task that any AI (ok, probably one that will exist a couple of years in our future) could compile and make a pretty presentation of in about a half an hour, while I have been working on it for many weeks. The process would be faster if all of the old info recorded on the originally handwritten museum catalogues for the Swedish Historical museum, and on their old typed card-catalogues had been entered into the database, in addition to being scanned. Much of that old information is, in fact, in their database, so if one clicks on an artefact on their web interface for their database, you get all the info. However, a number of the older items (especially those obtained in the 1800's) haven't yet gotten that information entered into the database. This means that if you click on an item, and it doesn't include anything in the way of a description or measurements, there is a chance that if you into the old version of the database web interface and click on the information symbol there, and look at the old scanned catalogues you might find out more:
http://kulturarvsdata.se/shm/object/362802 is from the older database, which opens the catalogue view https://catview.historiska.se/catview/index.jsp for that inventory number, if you click the little brown book symbol on the "more information" section of the first of those two links.
https://samlingar.shm.se/object/D042FB9F-A914-4155-934A-DAA6DA9451B7 is the link to the new database version for the same artefact as above.
Taking the time to check those old catalogues for each item is part of why this project takes so long. There is also the part about finding the coordinates for the location, so I can plot each artefact on the map, which is super easy for some items, as that is in the database, along with a link to information about the archaeological excavation that provided the artefact. Other times there is just a name of the province or, perhaps the parish, in which it comes, and then I have to look up coordinates for that.
Then there is making the figure, with all of the items for a given area shown as dots on the map, with photos of the artefacts themselves, and arrows to show which photo goes to which dot.
So, yah, not quite to the point where technology could do this for me, yet.
But I am very, very aware that the time is coming, and soon (we have the pieces--the "can write good paragraphs" and the "can read scans of 1800's handwriting", and the "can make good art using supplied photos" and the "can search multiple databases"), so it is just combining them, and persuading the AI to stick to the facts), and I know that this may well spell the end of this sort of thesis chapter for students, and so I wonder what next? What will students do when an easily report of every object of a specific material and time period, in every museum, everywhere, can be generated in seconds? Right now we are encouraged to focus and narrow the scope of our research to keep the project feasible in the time period available for it. That is going to be a big game changer, and I am really glad I have the chapter on geochemistry to give the thesis more value than my artefact summary will have when such things are created in seconds...
These things were all part of my dreams last night, none of the details of which stuck with me, but which left me with an overall feeling of redundancy.
http://kulturarvsdata.se/shm/object/362802 is from the older database, which opens the catalogue view https://catview.historiska.se/catview/index.jsp for that inventory number, if you click the little brown book symbol on the "more information" section of the first of those two links.
https://samlingar.shm.se/object/D042FB9F-A914-4155-934A-DAA6DA9451B7 is the link to the new database version for the same artefact as above.
Taking the time to check those old catalogues for each item is part of why this project takes so long. There is also the part about finding the coordinates for the location, so I can plot each artefact on the map, which is super easy for some items, as that is in the database, along with a link to information about the archaeological excavation that provided the artefact. Other times there is just a name of the province or, perhaps the parish, in which it comes, and then I have to look up coordinates for that.
Then there is making the figure, with all of the items for a given area shown as dots on the map, with photos of the artefacts themselves, and arrows to show which photo goes to which dot.
So, yah, not quite to the point where technology could do this for me, yet.
But I am very, very aware that the time is coming, and soon (we have the pieces--the "can write good paragraphs" and the "can read scans of 1800's handwriting", and the "can make good art using supplied photos" and the "can search multiple databases"), so it is just combining them, and persuading the AI to stick to the facts), and I know that this may well spell the end of this sort of thesis chapter for students, and so I wonder what next? What will students do when an easily report of every object of a specific material and time period, in every museum, everywhere, can be generated in seconds? Right now we are encouraged to focus and narrow the scope of our research to keep the project feasible in the time period available for it. That is going to be a big game changer, and I am really glad I have the chapter on geochemistry to give the thesis more value than my artefact summary will have when such things are created in seconds...
These things were all part of my dreams last night, none of the details of which stuck with me, but which left me with an overall feeling of redundancy.