Deepfake video conferencing fraud

These systems haven't even reached uncanny valley yet and they are already this convincing, a few more years as the technological singularity approaches and big corps will realise that they should have pushed sooner for some actual laws to be put in place restricting how this sort of stuff is used because by that point they will have already had to lose millions/billions due to human staff simply not knowing who they are dealing with.
 
A realtime deepfake! I've seen a few examples and they have been getting better, throw in a low quality video stream and I can see it fooling someone if done well enough.. but the examples have been a bit obvious.. however that is also part of the 'acting' and interactions that look wooden..
 
Mission Impossible 7, the Dead Reckoning which came out last year had a realtime deepfake on a comms device. It was during a chase sequence and the deepfake was convincing enough that it tricked some of the protagonists "go left! go right! go left!" etc, taking them off-route and putting the characters in danger.

This has become real.
 
Without seeing the conference it's impossible to pass any sort of judgement on the employee even though I have my doubts it looked legit because DF's still have an uncanny valley look about them despite improving
 
TBH it probably worked because of a bit of manipulation and people doing their job with little attention to detail or conscientiousness, etc. it probably wouldn't work, yet, on someone who did their job with a little diligence.
 
dowie said:
Fraud is already illegal.
Click to expand...
Yes but that doesn't stop lawmakers from bringing up the need to have legislation specifically regarding AI.

State lawmakers across the US are moving legislation that would place safeguards on artificial intelligence to curb deepfake videos on individuals,
Click to expand...
news.bloomberglaw.com

States Target AI Deepfakes in Taylor Swift Aftermath (Correct)

State lawmakers across the US are moving legislation that would place safeguards on artificial intelligence to curb deepfake videos on individuals, including recent ones featuring celebrities such as pop star Taylor Swift and President Joe Biden.
news.bloomberglaw.com news.bloomberglaw.com

What you're saying is applying a blanket category to "crime", like assault, but there are specific sub categories of assault carrying different penalties and so on. There is no AI sub category yet for this sort of thing.
 
Last edited:
I't a bit on the back of the new Biden niece fake and upcoming concerns for ww elections ;
I think they (meta/musk) should be able to indelibly mark known fakes as such, and leave them in circulation, so that peoplefools without due diligence, are aware they could be fooled.
 
StriderX said:
Yet millions of people do it every day so doesn't seem like anyone really cares to police it seriously.
Click to expand...

Millions of people do all sorts of crime every day... Must be because no one cares to police any crime. We should just draft a law that says "Everyone stop doing crimes please!"
 
mrk said:
Yes but that doesn't stop lawmakers from bringing up the need to have legislation specifically regarding AI.

[...]There is no AI sub category yet for this sort of thing.
Click to expand...

There's no sub category for cheques either in the fraud act 2006, it's still illegal to fake a cheque or steal one and forge a signature etc..

What are you asking for here, you're saying we need "actual laws" but what do you want these laws for specifically?

I'm not saying you're wrong to ask for some new legislation necessarily it's just unclear what you're asking for.
 
Last edited:
dowie said:
There's no sub category for cheques either in the fraud act 2006, it's still illegal to fake a cheque or steal one and forge a signature etc..

What are you asking for here, you're saying we need "actual laws" but what do you want these laws for specifically?

I'm not saying you're wrong to ask for some new legislation necessarily it's just unclear what you're asking for.
Click to expand...

Take about 20 percent off there bud. Not the fight you're looking for.
 
MoNkeE said:
Yes, Dowie would like an argument, please commit to a stance for him to counter-commit to :p
Click to expand...

No, I'm literally just asking for him to clarify. I don't disagree that legislation for AI nudes etc.. could be useful for example but that's rather different to fraud.

Like outlawing making pornographic images using someone's likeness is probably going to be easier to legislate for than outlawing making anything with someone's likeness as there are plenty of legal uses. It's like someone last century wanting to outlaw photocopiers or printers/scanners etc.. because they may be used for fraud.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom