News for hackers. Stuff that matters.
Re:Disney making $1B investment in OpenAI, will allow charac... by impulser_ on Thursday December 11, 2025 @10:01AM (#46232217)
I guess it will depend on how good their security is, because I'm assuming Disney is entering this deal with hard guidelines on what will be allowed.
Re:Disney making $1B investment in OpenAI, will allow charac... by edoceo on Thursday December 11, 2025 @10:41AM (#46232756)
AI with hard guidelines? I don't think that will work.
Re:Disney making $1B investment in OpenAI, will allow charac... by pjc50 on Thursday December 11, 2025 @11:37AM (#46233591)
"Hard guidelines" is making me think of the Pirate Code from Disney's Pirates of The Caribbean.
Re:Disney making $1B investment in OpenAI, will allow charac... by forgotaccount3 on Thursday December 11, 2025 @11:13AM (#46233214)
Not necessarily AI with 'hard guidelines' AI tools that pass output to a filter with 'hard guidelines' is definitely feasible.

Take the input as normal, pass it into Sora 2 and execute it as you would, pass the output through a filtering process that adheres to hard guidelines.

Of course, when talking about images, what is a 'hard guideline' here? Do you take the output and pass it through AI to identify if there's x y or z categorys of content here and then reject it?

Re:Disney making $1B investment in OpenAI, will allow charac... by bilbo0s on Thursday December 11, 2025 @10:16AM (#46232401)
This.

Don't believe for a second that Sora will allow you to make racist content with Disney characters.

That said, there are a lot of other models out there that care about neither licensing nor alignment. So those will allow you to make racist content. Then you can do whatever you like with that generated content.

A lot of IP owners will learn that there is more than one way to skin a cat. It's easier than people think to turn children's characters, like say, Hermoine, into a raging racist. And there's very little technically speaking that they can do to stop it.

But yes, on OpenAI specific properties, they can definitely stop it dead in its tracks. They can even get better at stopping it over time. In fact, the more users try to generate it, the better the system will get at stopping it.

Re:Disney making $1B investment in OpenAI, will allow charac... by zrobotics on Thursday December 11, 2025 @10:20AM (#46232464)
I'm sure that's what Disney's lawyers specified in the contracts and what their execs expect. However, judging by how LLM controls have gone in the past, I'm fully expecting to see a slew of awful content featuring Disney's characters in the days after this launches. OpenAI also probably won't ever be able to actually stop people from generating harmful content with the characters, but the volume of awful stuff will probably eventually slow down as people get bored and move onto the next controversial thing.
Re:Disney making $1B investment in OpenAI, will allow charac... by doug_durham on Thursday December 11, 2025 @11:57AM (#46233869)
I don't think that you will ever be able to generate Disney characters with the Sora app. Sora is both a model and an app. Instead I think that there will be a heavily guardrailed specialized app where you can do some highly restricted things for the opportunity for the content to show up on Disney+. Think of it as "Disney art! Powered by Sora".
Re:Disney making $1B investment in OpenAI, will allow charac... by dragonwriter on Thursday December 11, 2025 @11:43AM (#46233677)
> Don't believe for a second that Sora will allow you to make racist content with Disney characters.

Don’t believe for a minute that whatever filters it uses will be sensitive enough to the way racist content is constructed to stop people from doing just that.

Re:Disney making $1B investment in OpenAI, will allow charac... by mrguyorama on Thursday December 11, 2025 @01:51PM (#46235408)
"Don't let the user do racist shit" is literally AGI Hard territory of problem solving.

It's right up there with "Let kids communicate anonymously but not to perverts" and "Is this porn or educational?"

Re:Disney making $1B investment in OpenAI, will allow charac... by PunchyHamster on Thursday December 11, 2025 @10:33AM (#46232641)
> Don't believe for a second that Sora will allow you to make racist content with Disney characters.

Yes, because AI's so far have been oh so resilient to jailbreaks and oh so great at picking out the potentially "not aligned with corporate values" content...

Re:Disney making $1B investment in OpenAI, will allow charac... by corobo on Thursday December 11, 2025 @11:39AM (#46233618)
How are they going to stop it?

A certain combination of nonstandard characters will make an AI character drop an n-word no problem

I guess they could chuck the output through whisper or something to see if it transcribes back to anything dodgy?

LLM security feels very ball of sand held together with duct tape haha

Re:Disney making $1B investment in OpenAI, will allow charac... by dagmx on Thursday December 11, 2025 @10:20AM (#46232466)
How can you prevent it from making that content? People will find a way to combine certain words together to make it happen.
Re:Disney making $1B investment in OpenAI, will allow charac... by simonw on Thursday December 11, 2025 @10:20AM (#46232469)
Given the creativity of the jailbreaking community I will be very impressed if OpenAI manage to reliably prevent Sora from creating disagreeable content with Disney characters.
Re:Disney making $1B investment in OpenAI, will allow charac... by sarchertech on Thursday December 11, 2025 @10:39AM (#46232731)
If they’re jail breaking it, they should already be able to get it to create Disney characters without this deal.
Re:Disney making $1B investment in OpenAI, will allow charac... by andrew_lettuce on Thursday December 11, 2025 @11:06AM (#46233123)
Almost like this is not about the value of the deal, but the payoff to both sides from the headlines?
Re:Disney making $1B investment in OpenAI, will allow charac... by lm28469 on Thursday December 11, 2025 @10:34AM (#46232648)
> Don't believe for a second that Sora will allow you to make racist content with Disney characters.

It's not racist, it's an historically accurate depiction of 1930s Germany under the authority of a significant leader who may or may not be controversial today

Re:Disney making $1B investment in OpenAI, will allow charac... by butlike on Thursday December 11, 2025 @11:46AM (#46233712)
Great, so censorship by attrition.
Re:Disney making $1B investment in OpenAI, will allow charac... by conartist6 on Thursday December 11, 2025 @10:20AM (#46232460)
Once you make the content it's just content, no? How the hell are you going to ban racism? A lot of racism is "dog whistle" stuff anyway -- it's designed to convey a message to people who already know what is being suggested while seeming innocuous to any enforceable standards of decency... The racists will surely have no trouble bending the models to produce Disney content that is in practice used to promote and celebrate racism.
Re:Disney making $1B investment in OpenAI, will allow charac... by chaostheory on Thursday December 11, 2025 @10:21AM (#46232486)
Based on what I’ve seen on Sora, it’s likely possible but only for a day or a few hours. They’ve been getting better and faster at censoring.

I agree. Those characters are likely safe on Sora

Re:Disney making $1B investment in OpenAI, will allow charac... by andrew_lettuce on Thursday December 11, 2025 @11:08AM (#46233144)
That's still whack-a-mole though, just because they're rolling out new virus definitions quicker doesn't do anything to prevent zero day exploits