Tom Hanks says that’s not him selling dental insurance in Instagram ads
avclub.com
avclub.com
Edit: well it could be a face replacement over a video but not a deep fake one
This has to be intentional stylistic renderings.
Can't we do a similar thing where the actor/artist/whoever digitally signs the media they are in, the platform which is showing the media can then confirm the rights to the person.
Let's say Tom Hanks wanted to sell his likeness, that shouldn't stop him, he would just need to sign any media which uses it.
I'm sure somebody is working on this already, but who?
This is corny, but it summarizes the idea quite well. https://youtu.be/-gGLvg0n-uY
As I've always joked about conspiracy theories. It's only a conspiracy if you're not in on it.
Without a registry of people with unique names, what's to stop some other person named Tom Hanks from signing media?
(I still find AVClub unusable due to bad/LLM generated content though)
There are a ton of aging celebrities these days.
I'm not sure what point they are making with this run-on sentence. Fortunately, it's already against the law to use someone's likeness in this way without their permission.
The original proposal from the studios was to allow scanning likenesses of background actors in perpetuity, for a day’s wages, and then being able to use it at will without royalties. Which is a pretty good reason why they voted to strike.
At some point the software industry, of which I am actively a part, has to take responsibility for what we make. Hiding behind section 230 and/or "just following orders" (ahem, "requirements") no longer cuts it.
Professional ethics are demanded from accountants, lawyers, doctors, and physical engineering disciplines. There are real consequences in those professions when some fail to uphold the covenant expected of them.
Software engineering can no longer be solely about revenue without accountability.
The thing is, software engineering in regulated fields already is regulated. A degree is required to work on software for many medical device companies, regulations are in place to ensure code is written safely. Could all of this be improved? Sure. But your last sentence is just not true for regulated industry.
If you amended your comment to specifically address social media companies, then yes I agree.
If you run a self-service advertising system, then don't advertisers count as users, and don't their ads count as third-party content that the system operator isn't liable for?
They may not be your service's only users, and people may not like ads, and ads have a commercial purpose, and advertisers pay for them, but I don't think Section 230 cares about any of those things.
thank you for recognizing the connection I was making, you are a gentleperson and a scholar.
And at least IMO, what would the expectation be? That Facebook investigates every single ad to make sure that nobody in them is impersonating someone else? How would you do that? To what extent does Facebook need to fact check all of the details of all of the ads across its whole platform? That's exactly the sort of hassle that 230 was intended to prevent the need for in the first place.
Either figure it out or don't do it?
I hate advertising and I hate Facebook, but I think there's really nothing the platform can do to stop fake ads because—especially as deep fakes improve—there's nothing that any platform can do as an intermediary to practically prevent them.
That doesn't mean the advertisers can't be held liable (and they should). It's just wild to me to think that the intermediary should also be liable simply because they accepted money.
Same way that TV stations and magazines do.
YouTube has been promoting the exact same financial scam (Tessler AI investor created by Elon Musk) for years. The fraudsters haven't even bothered to change the video — there is ignorance, and then there is negligence. If platforms aren't liable for their advertisements in some respect, there's a direct financial incentive for them to look the other way.
Facebook should take a loss on this, how else are they encouraged to keep this off the site?
Edit: but just because you’re big and have automated most of your business, doesn’t mean you should be able to claim ignorance. If I build a robot that shoots enemy combatants and it walks around shooting civilians, I can’t just go ‘oh well auditing these things automatically is hard so I guess it’ll just have to work that way’.
What exactly is the timeline you're looking for here?
If you're making money at time X, you can be handling serious and personally-impacting problems at time X.
The media companies can then solve the risk like everybody else by requiring their ad providers to indemnify the media companies for any liability that arises from any lies in the provided ads. If the media companies do not trust the ad providers to have adequate funds to fully indemnify them then they can demand the ad providers to post bonds or insurance demonstrating adequate funds. Alternatively, the media companies can police their ads they display or demand adequate proof for high value ad placements to minimize their risk.
They have plenty of bog standard options to mitigate any strict liability risk. And assigning strict liability to the media companies for any lies guarantees any damaged parties full access to relief for any harms enacted on them.
That's ludicrous.
Therein lies the problem.
These are deterrences from reality, and if we were forced to take a close look at a large majority of online advertising we would plainly see that it ranges from the outright sketchy, to borderline lying and certainly user-harassing. They worry that if too much scrutiny were applied, they might actually face some consequences for their actions and we might reevaluate their position as arbiters of “what’s viable” on the internet.
Facebook just isn't one of those I guess.
Billions of not-ads get served on Facebook yet they were able to clamp down on stuff like ISIS uploading decapitations.
imagine if the FDA or some other sort of stringent agency had that kind of remark. "Well, we make 3 billion pills a day, of COURSE 15% of those are going to be deadly poison." -- wouldn't fly for a second.
the excuse of "Well it's going to be real tough to reduce the amount of poison pills while still making trillions in profit" would be equally useless; society wouldn't (and shouldn't) care about that.
A simple solution is requiring the companies providing the ads to be placed to indemnify the ad placers. That means the ad placers push the monetary liability onto their customers to only provide legitimate ads. As they know exactly who their customers are and can vet them it is trivial for them to go after them if they are sketchy.
Easy. Next problem.
"Oh gosh, I opened the door and a bunch of rats came in... I haven't closed it yet. What can I possibly do to take care of this rat problem?"
What a policing nightmare we've created...
Tom Hanks says that’s not him selling dental insurance in Instagram ads (https://news.ycombinator.com/item?id=37731670) (84 comments)