Comment by grey-area | Hacker News Reader

Comment by grey-area | Hacker News Reader

Parent Full thread

We can't reliably detect AI text, and they are getting better at producing plausible humanlike text (albeit still with flaws and mistakes). I do expect them to become impossible for humans to detect when the companies put effort into that (so far they've been focussed on other domains like dev or maths).

I'm afraid your dystopian future where the zone is flooded with shit and nobody can trust anything seems eminently plausible, partly because the internet was built without identity.

I think to solve this you'd have to have strong proofs of identity locked to biometric identifiers. We'll probably need that anyway for other reasons.

It's not a trivial problem. Websites like this could require webauthn tied to a hardware identifier of their choice (like a fingerprint or faceid) - I'm not sure if webauthn provides that level of detail to providers (which method was used), and how secure that attestation is. Of course that's not going to be popular but it might be the lesser of two evils.

> I think to solve this you'd have to have strong proofs of identity locked to biometric identifiers.

That's what surveillance-capitalism wants, but what regular people need is something which is easier to supply and less likely to be turned against them: A lesser proof of "I invested enough time/effort/money into this identity that I'll be hurt if its reputation craters".

Well it's what we need unfortunately, unless you want to talk to potentially malicious machines and think you're talking to a human. Identity is vitally important for us, the end users interacting with each other. Far more important than it is for nation states and companies, who have always had ways to track us and care more about us in the aggregate than as individuals.

I'm not sure I share your concern about surveillance capitalism - that ship sailed long ago - you are stored in many nation-state and corporate databases under all the aliases, phone numbers, IPs you've ever used. I sincerely doubt your opsec has always been even near good enough to evade this surveillance, it's a little late to start worrying about it now, particularly when there are other concerns.

I'd be more concerned about billionaires spending on bots impersonating humans to sway elections in my local state than a nation state snooping on forum comments and linking them with something else.

Proof of identity is badly needed in order to use the online spaces we've come to know and trust, and time is running out.

> Well it's what we need unfortunately, unless you want to talk to potentially malicious machines and think you're talking to a human.

If my options are "give all my data to malicious corporations and their bots" or "wonder if the pathological liar I'm talking with is a bot" I'll stick with guessing about the liar. Proof of identity is a scam, not a solution.

We shouldn't abandon the entire idea of an internet where people can find information and speak to each other without being identified and monitored 100% of the time just because the companies who want us to be identified and monitored filled the web with bullshitting bots

Maybe you've already given up and no longer care about that freedom, but don't count on everyone else giving up theirs.

You are dreaming if you think that has been the reality for the last 20 years. Your data is all sitting on servers in Utah, either decrypted or waiting to be decrypted. You can easily be identified and tracked by your online footprint, unless you take really quite extraordinary measures.

The internet does not provide anything close to anonymity and never really has. Even a simple IP address is enough to out the majority of users, when combined with fingerprinting, metadata tracing and the use of metadata from other users, there really is no possibility of hiding your identity from a determined attacker.

I would choose the opposite option personally, and would welcome it if this site for example allowed biometric identification and used that to weight comments. While I wouldn't want to force that on anyone I do think it's going to be necessary given the rise of LLMs.

> there really is no possibility of hiding your identity from a determined attacker.

If there's no possible way to hide your identity then as long as we're determined to identify humans and bots we already can. In truth, most people are able to hide their identity pretty well from most people most of the time. The government can pretty much always track down who you are if they feel like it. If you're using a cell phone apple and google can find out who you are. If you use windows Microsoft will have a pretty good idea of who you are, but the vast majority of websites can't get anything more identifying than an IP address and a user-agent unless you tell it to them without trying to correlate your traffic with other traffic. That's where browser fingerprinting comes in really, it tells them the same browser a person used to log in with their name and address at one website is using another website. That lets them make some educated guesses, but they can't be sure who is behind the screen and there are easy steps a person can take to make it even harder for them to guess.

The fact that anonymity on the internet isn't perfect doesn't make it useless. Giving that up because you're afraid of reading something a bot wrote seems crazy.

Well, surveillance capitalism now has a killer app.

Utilizing terms like surveillance capitalism, doesn’t serve as a useful deterrent, because the underlying system has changed.

Previously Bots were piloted to be marginally effective at appearing to be normal users. With LLMs, that’s basically out of the window.

At this point your false positive detection rate is going to start hitting human users.

Users are also going to accept AI content. Simply because their feeds are going to provide it to them.

It’s cheap, and it achieves the same dopamine hits that human interaction will achieve.

Unless you are knowledgeable in a field, you wouldn’t be able to make out if something is off.

The most likely positive outcome is we see bots that take on bridging roles in conversations.

The question I would ask is why would and who would, pay to verify a user is human ?

> Unless you are knowledgeable in a field, you wouldn’t be able to make out if something is off.

Maybe in the future that will be true, but right now AI is constantly making very obvious mistakes and errors. That might even eventually become one way we tell if something was written by AI. If human pathological liars get flagged as AI by mistake, that wouldn't even be so bad.

Reply on news.ycombinator.com