This is a sorites problem and you want to sort some pebbles into kinds. You've made clear that externalism about porn may be true (something may begin or stop being porn in virtue of properties outside it's own inherent content, such as where it is and contextual features).
It seems to me that you have to prioritize your goals in this case. So the goal "ban porn" is much more important then the goal "leave eroticism alone". My response would be to play safe and ban all footage including genitals, similar to what the Japanese already do with pubes etc...
This response is analogous to your Oracle AI paper suggestion, only mine is less sophisticated.
Now, I can steelman your request, you must ban all porn and with the exact same importance you must endorse that eroticism take place. Then I'd evoke the works of skeptical philosophers about the sorites problem, like Paul Unger, who doesn't think people nor himself exist (while still thinking you should not "live high and let die").
I'd argue that these theorists are right, and that thus there is no matter of fact as to where porn stops and erotism begins.
The goal is still there though, the goal is embedded in Nature, regardless of what we think about it? Ok. Now if we know there is no fact of the matter, and erring in both directions would be equally bad, we should make pools to find out the general opinion on the porn eroticism divide (seems safe to bet that porn is porn in virtue of its relation to human minds, since chimp porn is not attractive to us). Those pools would be compulsory, like voting is in some places, we'd have as good an idea about what the human mind considers porn as we can. Like hot or not, there would be worldwide porn or not websites.
This would go on (causing a lot of damage every time someone had to detect porn) for a while, until the best algorithms of the time could do the detecting indistinguishably from the average of all humans to some approximation. When that was done, anyone attempting to create eroticism would have their files automatically scanned for porn-ness by the narrow AI.
We would have the best system possible to avoid infractions in both directions of the sorites problem. For every doubling of the human population from then on, re-check to see if our senses have shifted drastically from the previous ones and adapt the algorithm.
Seems desirable to over the long run make sure cultural globalization of morals is very intense, so I'd ban immigration laws to make sure humanity clusters around a narrow subset of mindspace as it regards porn or not.
Depending on how that goal factors in with every single other goal we have ever had, it may be good to destroy lots of resources after the moral globalization process is done, so that people spend more resources on survival and have fewer chances of drifting morally apart. Following the same reasoning, it may be desirable to progressively eliminate males, since males are less biologically necessary, and porn judgement varies significantly by gender.
For every doubling of the human population from then on, re-check to see if our senses have shifted drastically from the previous ones and adapt the algorithm.
I feel this kind of idea could have some AI potential in some form or other. Let me think about it...
To construct a friendly AI, you need to be able to make vague concepts crystal clear, cutting reality at the joints when those joints are obscure and fractal - and them implement a system that implements that cut.
There are lots of suggestions on how to do this, and a lot of work in the area. But having been over the same turf again and again, it's possible we've got a bit stuck in a rut. So to generate new suggestions, I'm proposing that we look at a vaguely analogous but distinctly different question: how would you ban porn?
Suppose you're put in change of some government and/or legal system, and you need to ban pornography, and see that the ban is implemented. Pornography is the problem, not eroticism. So a lonely lower-class guy wanking off to "Fuck Slaves of the Caribbean XIV" in a Pussycat Theatre is completely off. But a middle-class couple experiencing a delicious frisson when they see a nude version of "Pirates of Penzance" at the Met is perfectly fine - commendable, even.
The distinction between the two case is certainly not easy to spell out, and many are reduced to saying the equivalent of "I know it when I see it" when defining pornography. In terms of AI, this is equivalent with "value loading": refining the AI's values through interactions with human decision makers, who answer questions about edge cases and examples and serve as "learned judges" for the AI's concepts. But suppose that approach was not available to you - what methods would you implement to distinguish between pornography and eroticism, and ban one but not the other? Sufficiently clear that a scriptwriter would know exactly what they need to cut or add to a movie in order to move it from one category to the other? What if the nude "Pirates of of Penzance" was at a Pussycat Theatre and "Fuck Slaves of the Caribbean XIV" was at the Met?
To get maximal creativity, it's best to ignore the ultimate aim of the exercise (to find inspirations for methods that could be adapted to AI) and just focus on the problem itself. Is it even possible to get a reasonable solution to this question - a question much simpler than designing a FAI?