Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

India ruling party's IT cell used AI to show smile on arrested protesters' faces

altnews.in

Post reply on HN
Print view View original

441–450 of 478 posts

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#441

Post by tacocataco »

joshspankit wrote: post #114

How will we verify photos and videos without dissolving in to endless speculation? I’m genuinely asking as this is something I’ve been worried about for a while. The only thing I can think of is storing a hash or “phash” with a trusted third party that can then be used to later verify which image is the original. For video, store the hash/phash of the frames with timestamps. But would that even work? How do you back-…

Stop trying to verify photos and ask the people that were there when the photo was taken?

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#442

Post by tacocataco »

adrr wrote: post #257
ilrwbwrkhv wrote: post #37

What's with India and the shift towards the right? I thought it had too many religions and different cultures to pull this off.

World has moved towards authoritarian governments. Haven't seen a good reason why this push but there seems to be correlated to the amount information people have access to(eg: social networks, digital media etc).

Wasn't there plenty of authoritarian governments before the internet?

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#443

Post by deno »

anigbrowl wrote: post #371
deno wrote: post #226

Earlier quoted context omitted.

I see this as completely different issues. Whether AI exists or not this is and was a problem. I’m proposing technical solution because AI fakes are a technical problem. And yeah if people get used to green checkmark you can’t just pump fakes willy nilly, you actually have to put some effort, as before AI. I guess the problem I want to solve is having to live in a world where you have to assume absolutely everything…

First you have to solve the collect action problem of getting everyone to care about the green checkmark, then you have to solve the new technical problem of just making photographs of high-quality altered images.

I fully acknowledge the second issue, and have offered some potential solutions in my other comments, but I think this can have tremendous impact even if you don’t solve the second issue in first draft.

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#445

Post by GoblinSlayer »

AnimalMuppet wrote: post #58
pjc50 wrote: post #38

Earlier quoted context omitted.

That is what a lot of us think the real threat of "AI" is: not that it will autonomously produce bad outcomes for humanity, but because it's a tool that's indifferent to truth or falsehood it allows any kind of truth or reality based politics to be totally flooded out of the public sphere. Everyone ends up fighting in the hall of mirrors.

Yeah. The postmodernists (as I understand them) have been talking about "post truth" and "claims of truth are just assertions of power". And I've always thought that their position was insane. Now they look more like prophets. That world is increasingly the world we live in. The question becomes, how can we stay sane in that world? How can we find enough actual truth (truth that is actually true , true in the old sen…

>The question becomes, how can we stay sane in that world?

Diversify your information sources.

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#446

Post by joshspankit »

EVa5I7bHFq9mnYK wrote: post #423
joshspankit wrote: post #389

Earlier quoted context omitted.

> You can apply a filter and immediately after compute a hash. Can you expand on that? I’m not sure what wouldn’t work

>> For video, store the hash/phash of the frames with timestamps. But would that even work? That would not work, because how do you prove that hash has been computed before video manipulation? Already, the imaging chips do some processing, so you never get the true raw image from the CMOS sensor.

Yes, I agree. I don’t think that method could be used if the manipulator was the person taking the photo or the one doing the initial processing. I’m (personally) less concerned about the true raw image vs lossy compression (i.e. JPG), and more concerned with things like the OP where an existing image was changed to show detainees smiling and that image was then used to manipulate the public.

The way it is now, it could easily have turned in to a “he said, she said” situation with each side claiming that their pic is the real one and the other’s was manipulated. In a case like that we could (theoretically) look back in the record to see which hash was recorded first.

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#447

Post by joshspankit »

tacocataco wrote: post #441
joshspankit wrote: post #114

How will we verify photos and videos without dissolving in to endless speculation? I’m genuinely asking as this is something I’ve been worried about for a while. The only thing I can think of is storing a hash or “phash” with a trusted third party that can then be used to later verify which image is the original. For video, store the hash/phash of the frames with timestamps. But would that even work? How do you back-…

Stop trying to verify photos and ask the people that were there when the photo was taken?

I don’t think that’s going to be enough. How do you verify that it’s them who is responding, or that they are not being coerced? (before anyone comments, yes I know I’m taking this to the extreme, but if we handle the extreme cases we handle the easier ones too)

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#448

Post by bogwog »

carlitossway wrote: post #307
bogwog wrote: post #200

Earlier quoted context omitted.

That doesn't seem too unbelievable. Maybe a sufficiently large LLM will be able to find a flaw in the algorithms or logic behind some encryption schemes which even our greatest monkey brains overlooked.

No, they cannot. LLMs are LANGUAGE models, not KNOWLDEGE models. They do not "know" what algorithms mean any more than they know what everything else they "talk about" mean. What they are more than capable of is to produce really convincing hype, which claims that LLMs can do anything and everything.

Some skepticism is healthy, but I think you have a misconception about what an LLM is. Neural networks use machine learning to automagically find connections between data. Handwriting recognition is an example of this, as it's trivial to build a model that can recognize the differences between characters/digits, whereas writing an algorithm to do that effectively would be very difficult. You can train a model on sample data, and it will "learn" the patterns that make up a '7', 'G', 'O', etc with the same or better accuracy than a person.

Language models are more or less the same thing, except they're trained on human language. Since human language is an encoding of human thought processes, then the model learns the patterns of human thought which are embedded in text, and can both generate and recognize them.

This is why LLMs can "understand" and answer complicated questions that don't exist in their training set, and do things like debug code. Even though it never saw that exact question, it recognizes the high level/abstract concepts that make up the question (since those abstract concepts exist in various different forms in the training set).

So giving an LLM a textual description of an algorithm, and having it detect logical errors in it, is not that unbelievable. Whether current models are good enough to do it well is the question to ask.

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#449

Post by gofreddygo »

pjc50 wrote: post #38
juujian wrote: post #6

That is getting real freaking dystopian. Sure, we would have been able to do this before, and the Soviet Union did a ton of photo editing. But doing this particular edit must have been so easy. No cost at all, no mental barrier. You would almost think 'why bother'. And before anyone says 'yeah, but the truth came out'. Well, we saw it, but how many of the particular audience of whatever that outlet was will never see…

That is what a lot of us think the real threat of "AI" is: not that it will autonomously produce bad outcomes for humanity, but because it's a tool that's indifferent to truth or falsehood it allows any kind of truth or reality based politics to be totally flooded out of the public sphere. Everyone ends up fighting in the hall of mirrors.

The risk, danger and shallowness of perpetrating lies supported by fake images does not stop at politics.

Think fake instagram accounts with images, coercing and shaming victims for personal revenge. Shady companies selling fake stories supported by fake pictures of "satisfied" users. Influencers and celebrities making themselves look younger, prettier, fuller.

you see anyone who wants to spin a web of lies, AI as a tool makes it easier, possible and in som cases unavoidable.

Re: India ruling party's IT cell used AI to show smile on arrested protesters' faces

#450

Post by sriram_malhar »

codegeek wrote: post #133
dagmx wrote: post #22

As an NRI, I continue to be disappointed by the country I called home for most of my life. Between the politics and the pollution in Delhi, I fear I can never really return home. The last time I was home, I was struggling every day to breathe because Delhi has crazy air in the winter. While high school friends were telling me about how they don’t believe in equal rights, justifying any means to achieving a Hindu coun…

For non Indians, NRI = Non Resident Indian i.e. an Indian who lives outside India. In case someone wondered. It is easy to shit on the party in office which is BJP right now but show me a country that can have hundreds of languages, all religions and still survive for centuries as one country. Europe couldn't do it. They broke into EU (for better or worse). There are flaws but India is a complex country of various cu…

> still survive for centuries as one country

You have drunk the koolaid, my friend.

India became 'India' in 1947. Before that (and even now) it was exactly like Europe ... many languages and cultures and food very very different from each other (ask a Naga what he has in common with Tamilians or Kashmiris), a large shared Hindu ethos similar to how Christianity was a common factor for Europe) and lots of warring kings with their own territories. There was no "Indian" identity.

About Congress ... no defence about their fuckups. But don't be so dismissive about what they accomplished. I say this as no fan of Indira Gandhi.

Remember that India had become free after 200 years of enslavement and harm to our self-image, from being one of the world's richest areas to the poorest, and millions of lives had been torn asunder under partition.

In Congress's era, esp. Jawaharlal Nehru's time, we got a beautiful constitution, the premier institutions and facilities (TIFR, IISc, IITs, FTII, BARC, AIIMS), developed a massive power infrastructure with hydel and nuclear, became a nuclear strong power, developed an in-house rocket and satellite program etc. It had a strong court system; recall that Indra Gandhi was thrown out by the Allahabad supreme court on electoral fixing.

Manmohan Singh was (and is) a learned economist who righted many of IG's fuckups. He led the 1991 reforms. A large large portion of the projects that Modi has lent his beaming face to were fully funded and started in the Manmohan era (Chandrayaan, Mangalyaan, Atal tunnel, UPI, Aadhar, so many airports and highways). The Modi govt has been successful in rebranding all of them in his image, but don't be fooled.

I can give you a huge laundry list of things that are terrible under Modi. The COVID response, the complete takeover of august institutions by deranged Hindus, the complete submission of the media, the complete takeover of massive projects by a few crony friends (a profitable and well-run Mumbai airport was taken over by Adani, who literally had zero experience in running airports, with no formal publicly transparent process). The "PM Cares" fund, which every single govt employee and every corporate was FORCED to hand over money too, apparently is not a fund of the govt; there is no transparency about how much money has been collected, let alone spent where. I could go on about how every minority (Dalits, Muslims, tribals) have been rogered.

Post reply on HN