-7

Honest question. There is no known photo of Homer Plessy, of Plessy v. Ferguson. The photo which often turns up in searches is absolutely not him; a case of mistaken identity now well documented.

I have created an A.I. photo based on descriptions of Homer Plessy of the time, and presumably from scans of his descendants.

I'm clearly labeling it as A.I., but is generating such a picture playing with fire? He was a real historic person, and this photo is conjecture.

Is this immoral? Why or why not?

4
  • 2
    What do you mean by "presumably from scans of his descendants"? What leads you to think the image is based on such scans? Commented Oct 4, 2023 at 5:50
  • 2
    I'm not sure that "morality" is properly within the concerns of history. Morality should be discussed with moral authorities, using the sources and methods appropriate - your cleric, or a stack exchange appropriate to your faith, or even philosophy. What are the historical implications? @pere seems to have an insight below, that constructed images have a longstanding position within the study of history. What makes this image different from those (from the perspective of historical scholarship)? Commented Oct 4, 2023 at 11:18
  • 2
    I know I've been guilty of talking in moral terms here before, but I try to save that for extreme (and IMHO incontrovertible) cases. Having expertise in history doesn't really qualify our userbase for being experts in morality, so I'm thinking you really mean to be asking something else? Do you want to know if site policy does/should allow this? Do you want to know if we think its a violation of the StackExchange Code of Conduct? I can write on those subjects here, but otherwise this looks to be off-topic. Commented Oct 4, 2023 at 13:23
  • 1
    I've removed the fake/hoax picture itself. Its not nessecary to see it to answer the question, and a flag rightly pointed out that leaving it up generated hits and cred for a fake. If anyone really needs to get a look at it, check out the question's edit history. Commented Mar 25, 2024 at 16:05

3 Answers 3

Reset to default
13

Yes. It is immoral to create a fake photo of an historical person.

History is about facts. This is fiction. This is a hoax, a fraud, and should be immediately destroyed and removed from the internet. Removed from this site.

Historians are in a war every day to prevent lies from being established on the internet as 'truth'. AI generates lies that are difficult for the common person to distinguish from the truth. The very existence of this fake image on this site has already been absorbed by the search engines as 'fact'. But its a lie, a fraud, a hoax. It degrades the very nature of the world we live in by filling it with more lies. We have enough politicians and trolls trying to do this already...

So yes, creating a fraudulent image passing it off as a historical person is immoral.

(I can not speak to the accuracy of this image or its utility as a possible 'artists impression' since we have no information concerning the tools used to create it.)

Images generated by AI art are fascinating, and can be quite interesting, but the result is completely based on the model being used and the prompt being fed into that model. It in no way will perform any reasoning, thought, or artistry to create an image, it is merely a reorganization of material it has viewed during the AI models training process. Note the two watch chains on this image-it copied from two existing images somewhere, each of which had a watch chain in different locations, so without proper negative prompting it generated two watch chains in this image. This reiterates the copycat nature of this tool.

As I stated above, the danger here is that this image, like many AI generated answers, looks close enough to be considered valid. In these days of web scrapers gathering information, search engines fighting to introduce AIs which not only perform searches, but try to present answers to queries placed in their search bar, it is inevitable that, even if the OP clearly states the image is not real, it will be taken out of context and presented as fact by a search engine somewhere, and someone will believe it.

(As a side note, I will point out that I am not speaking from complete ignorance. I have 'tinkered' with AI art generation for in excess of 100 hours, generated hundreds of images, so have at least a basic understanding (I hope) of the processes involved.)

4
  • 4
    It is unhistorical, but it doesn't make it immoral - unless it is not clearly marked as fiction/AI and may mislead somebody. Reconstructions of this kind are done all the time - e.g., for pre-historical people, dinosaures, etc. What may pose mote reason for concern here is the privacy rights of this person. Commented Oct 4, 2023 at 5:59
  • The OP defined the terminology to be used. Commented Oct 4, 2023 at 11:16
  • 1
    @RogerVadim Plessy died in 1925 – his rights to privacy have expired. But if the picture was made using pictures of living (or recently deceased) relatives, their rights might be implied. Commented Oct 4, 2023 at 12:45
  • @ccprog I didn't literally mean his legal rights - the question is about morality. E.g., it can be perceived as abusing the memory of this person. E.g., like posthumous baptism of holocaust victims Commented Oct 4, 2023 at 13:22
8

Artists' impressions of historical people and events are ubiquitous, often based on very little of no factual information. You can see a lot of illustrations and statues of medieval and older kings whose likeness is absolutely unknown, and catholic and orthodox churches are full with conventional images of saints from nearly 2000 years ago. Nothing of that is immoral because the viewer knows that the image is just imaginary and only conveys the artist's ideas about the subject - which can actually be quite useful in its context.

Therefore, it's not immoral to post such an AI generated image as long as it's made clear to the viewer that it's just imaginary.

That problem doesn't show up just in history: it is very common in astronomy, and serious publishers of popular science are very consistent in labelling as "artist's impression" any image that could be misunderstood by an uneducated reader as a photograph.

Another question is whether such an image is useful, and I can't imagine any reasonable use of your image of Homer Plessy in History.SE, and nearly none outside of it.

2
  • I think there's a difference here, in that its almost never possible after the fact to double-check the AI's sources or quiz it about why it made the decisions it made. If we can't find out what it made up, we have no choice but to assume the whole dang thing is made up. Commented Oct 5, 2023 at 14:12
  • As an example, take the picture presented. Its supposedly a picture of an African-American. However, the photo looks very much like a white man. What we don't know is if the AI found a source that indicated he was in fact rather light-skinned, or if the AI just stupidly did a generic "picture of man from 1800's", and it looks like a white dude because most men in pictures back then that it found were white dudes. Commented Oct 5, 2023 at 14:22
0

I think the only time I'd ever want to see a non-genuine image of a historical figure on this site would be if its the subject of a "Is this actually what X looked like?" question (an question), and the image in question is actually being circulated elsewhere under the pretense of being an image of X. Drawings made up on the spot don't count.

That rules out an AI generated image in almost all cases. AI's are known to make up information, and cannot be quizzed by third parties as to the rationale and sources behind the decisions they make.

If an image of unknown provience used in an question turned out to be AI generated, that information alone would be enough to answer the question in the negative. If that information was known when the question was asked, the question should be closed as trivial/general reference.

2
  • Not sure the new user is going to follow their migrated question here, so I'm answering it as if it was a quasi-site-policy question. Commented Oct 5, 2023 at 14:10
  • Looks like your comment in chat last week about setting a policy concerning basing questions on ChatGPT output may need to be considered/established. Yet another question here Commented Oct 5, 2023 at 21:18

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.