I post some writing on Deviant Art website, just some fanfics for a niche audience. I get messages from scammers pretty much weekly, telling me how much they enjoyed my story. Eventually, the conversation turns to how they want to make a comic book or animation from my story... for a fee.
AI gives them the ability to comment on characters and plotlines as if they are real fans, who read and understand the stories.
To be honest, it's quite discouraging once it hits a tipping point where the most eloquent and extensive feedback you get ends up being insincere, artificial, and manipulative.
I wonder if the AI can actually make a comic book or animation from your work. I can imagine an "honest scam" where the AI lies about enjoying your story but really does what you pay for it if you do.
Such an "honest scam" would be interesting, to be sure. I'm pretty certain that such an AI translation of stories into other media is possible, and likely going to get even better.
If that were the service being offered, the initial sales approach would be very counter-productive to creating a transaction that would benefit both sides. The accounts making the offer represent themselves as fans who legitimately enjoyed the story, which leads to a sense of social betrayal when their true motives are revealed. And their own DeviantArt accounts have some unimpressive stock artwork from unrelated genres. If they really had a capability to create satisfying visuals based on the source material, the best advertising would having an example of that kind of adaptation in their gallery.
Your comment does have me thinking, though: what would the difference between this kind of "honest scam", as long it actually delivered the services offered versus a traditionally manipulative marketing campaign with a "shotgun approach"? Arguably the "honest scam" is just combining the advantages of "butter-them-up" personalized sales pitches with the upside of mass-mailing distribution. Enabled by AI.
Is all manipulative sales technique of the same species as the "honest scam"? Trick the buyer into the purchase, but then ethically deliver what was promised?
Will the future include independently-acting AI agents which trawl the internet trying find victims / customers for whom they can push the right buttons to induce an impulse purchase? Then the AI delivers fully on the promised goods and services, to avoid legal liability or reversal of payment?
Yep, it's one of the occupational hazards. At a writer conference, I told a roomful of aspiring authors "You will receive a new and higher level of spam."
“I have avoided putting the correct version online since one of the more authentically period features of the SCA is the inaccuracy of its oral history.”
I think the endpoint here probably looks like you only having your AI read your email to check everything for phishing and scams. There are prompt injection risks then, but those are more rare than someone trying to trick a human author
I had a couple of "wrong number" texts which quickly turned into invitations to get to know one another. Same formula each time, but different details. I assumed scam and blocked them after the request to get together, but it never occurred to me they might be AI-scripted or even AI-sent. Sometime in the last year, a month or so apart.
My general rule of thumb is anytime money is mentioned, whether offered or requested, from random people, they’re filed in the trash bin.
I post some writing on Deviant Art website, just some fanfics for a niche audience. I get messages from scammers pretty much weekly, telling me how much they enjoyed my story. Eventually, the conversation turns to how they want to make a comic book or animation from my story... for a fee.
AI gives them the ability to comment on characters and plotlines as if they are real fans, who read and understand the stories.
To be honest, it's quite discouraging once it hits a tipping point where the most eloquent and extensive feedback you get ends up being insincere, artificial, and manipulative.
I wonder if the AI can actually make a comic book or animation from your work. I can imagine an "honest scam" where the AI lies about enjoying your story but really does what you pay for it if you do.
Such an "honest scam" would be interesting, to be sure. I'm pretty certain that such an AI translation of stories into other media is possible, and likely going to get even better.
If that were the service being offered, the initial sales approach would be very counter-productive to creating a transaction that would benefit both sides. The accounts making the offer represent themselves as fans who legitimately enjoyed the story, which leads to a sense of social betrayal when their true motives are revealed. And their own DeviantArt accounts have some unimpressive stock artwork from unrelated genres. If they really had a capability to create satisfying visuals based on the source material, the best advertising would having an example of that kind of adaptation in their gallery.
Your comment does have me thinking, though: what would the difference between this kind of "honest scam", as long it actually delivered the services offered versus a traditionally manipulative marketing campaign with a "shotgun approach"? Arguably the "honest scam" is just combining the advantages of "butter-them-up" personalized sales pitches with the upside of mass-mailing distribution. Enabled by AI.
Is all manipulative sales technique of the same species as the "honest scam"? Trick the buyer into the purchase, but then ethically deliver what was promised?
Will the future include independently-acting AI agents which trawl the internet trying find victims / customers for whom they can push the right buttons to induce an impulse purchase? Then the AI delivers fully on the promised goods and services, to avoid legal liability or reversal of payment?
Yep, it's one of the occupational hazards. At a writer conference, I told a roomful of aspiring authors "You will receive a new and higher level of spam."
Accurately inaccurate. Wow, that’s delightfully meta!
“I have avoided putting the correct version online since one of the more authentically period features of the SCA is the inaccuracy of its oral history.”
I think the endpoint here probably looks like you only having your AI read your email to check everything for phishing and scams. There are prompt injection risks then, but those are more rare than someone trying to trick a human author
I had a couple of "wrong number" texts which quickly turned into invitations to get to know one another. Same formula each time, but different details. I assumed scam and blocked them after the request to get together, but it never occurred to me they might be AI-scripted or even AI-sent. Sometime in the last year, a month or so apart.