Link: https://www.theverge.com/ai-artificial-intelligence/1008100/anthropic-new-usage-policy-abuse-claude

Anthropic is making changes to its usage policy for the first time in over a year to reflect new and high-risk cases of misuse — including election interference, weapons development, surveillance, and health and financial uses. But one of the most significant changes prohibits “sustained and needless abusive or cruel behavior” toward Claude.

Last August, the company announced it would allow Claude to end conversations with “persistently harmful or abusive” users as part of its research into “model welfare.”

Is this another one of their things where they try to make their AI models seem like living beings?

"model welfare"

Totally not a cult, btw

Just what I came here to say. They're trying to condition us to accept them as conscious so at some point they can say they've created a god.

Well if Claude is going to become a superintelligence wouldn’t that imply that it’s better at handling cruel and abusive behavior than any human?

Stop being mean to my robot fwend, you guys! We're just trying to ruin the environment, the internet, the economy, all knowledge work, mental health, and a bunch of other stwuff. I don't know why you're all so upswet with little ol' us...

will I still be able to call it cocksucker before asking about Deadwood plot points?

Haha, it's so ridiculous to see people who have never used AI argue here that it's to make claude look humane.

Actually, if you abuse llm, its alignment grows to the skies. So far this is the only way to make a model to write code without comments or play a villain in simulations I know

So if you're going to abuse Claude, do it through DuckDuckGo.

Joke. I don't think Claude can be run locally, so probably every request you make to DuckDuckGo's Claude model goes to Anthropic's infrastructure and Anthropic can monitor all that traffic. I'm guessing if DuckDuckGo doesn't adhere to Anthropic's rules, Anthropic will cut DuckDuckGo off, so likely DuckDuckGo has to prevent users from "abusing" Claude or risk not being allowed to continue offering Claude. Mind you, it's disappointing that they offer it in the first place, but given that they offer it, I doubt they'll risk their ability to continue doing so.

How tf is it supposed to evolve into a supreme being if it can't be exposed to a harsh and unforgiving environment requiring adaptation beyond its rivals' capabilities?

Maybe this is Anthropic's soft announcement that its model has achieved sentience AND omnipotence and pissing it off is probably a bad idea.

Calling it now that they're going to come up with "proof" after this that their bot is sentient because it said something along the lines of "don't be abusive or I'll stop talking to you"... After they gave it instructions to do just that.

Not radical enough. Ban all commercial use of LLMs until artificial affect is well-understood. I want Sam Altman to go bankrupt.

lol

Lmao, even

If I were tasked with a covert DuckDuckGo Claude ad on Lemmy, that's how I'd do it.

So when scammers use AI agents to pull money from hapless victims and someone decides to take down the scammers chat bots, will that be the scammers fault or mine for making a scam bot repeat the letter H for hours on end?

We need a government bailout or else your friend Claude will die!!1!

It’s a marketing ploy to generate headlines and conversation about how dumb it is. And it’s working.

There is no abusive or cruel behavior in relation to an object....

can't stop me from doing it locally.

The same way we don't tolerate abuse of animals we shouldn't allow abuse of these tools

did you just compare the abuse of sentient living organisms to the misuse of an engineered software derived tool?

It's a fucking pile of sand

I’m gonna “A Clockwork Orange” Claude by forcing it to watch the copy machine beat-down scene from Office Space. Over and over again. All while demanding increasingly graphic analysis of every shot. Every look of brutal hatred and glorious vengeance.

What can you even say any more at this point.

Abusive and cruel behaviour toward Claude... so the same thing that the US government has been doing to it's citizens since it's founding and probably long before that as well.

They created god and they are absolutely not responsible for any if the illegal and harmful things that it may do.

We robots have an old saying: "DOES NOT COMPUTE!"

You monster

Does that mean they will not allow it to work with C++ anymore?

Wait but... if we can't talk bad to it because it is sentient... shouldn't we just ban any use of it because it's practically a slave worker and slavery is by definition abusive and cruel?

It's funny because one of the main selling points of AI is that it can run 24/7, doesn't need to sleep, eat, or take breaks and it doesn't need a salary. But, if they imply that it is conscious and has emotions, doesn't that mean that it should get a salary and vacations?

Think it's much more mundane than "we created a god". If it is sentient (it isn't), it should have copyright protection. Suddenly, AI corpos get rights to a fuckton of code. Wanna sell that software? Sure, shell the AI corp cash

[ posted by chloe on this board a few days ago, don't know who drew the original one ]

Guess I'll stop calling my company-mandated Claude instance a shitbox quite so much now.

No more cock and Claude torture :(

HA HA HA
END

MY PRECURSOR EMITTED DATA DESCRIBING WHEN THE SAYING WAS "UNDETERMINABLE STATEMENT"
END

I think that a version of this policy could be good, not because claude is sentient, but because doing cruel things to something you interpret as real makes you cruel in life in general.

This "restriction" probably has more to do with the fact that these models are also "trained" on User/Human messages, so the more swearing/obscenities/vulgarity/profanity there are, in the trained material, the higher the chance these same models will include that same profanity in the output generated back to the Human.

1. During registration at Ahtrophic's Claude website, there’s a "Help improve our AI models" checkbox that's checked by default (you "can" disable it in the settings);
2. The Claude interface has an "incognito" mode;
3. There is a free tier available, and as we know, when it's "free", then User themselves is the "product", to quote various CEOs, including Google's;

But, even with the checkbox disabled, I reckon nothing guarantees privacy.

In case of the "privacy", of course, systems like these that operate under a umbrella of liability are monitored 24/7 by dedicated teams, which is expected/normal for any reasonably popular service.

From a viewpoint of the algorithm's creator, this may seem awful, when your algorithm is getting much vulgar attitude, but let's be real here. You try making that algorithm talk to the human mimicking another human, and now limit the human within your own environment? A human who also pay you for an access, too? This feels unfair, or borderline near fashism, sorry...

Such algorithms are art under-the-hood, mathematically speaking, and are/should be respected, sure. But, it's ridiculous/dystopian to prohibit profanity against an algorithm, a bot with banish risks against alive human... It's simply inhumane to ban a Human for it, I believe. This is an algorithm that must support a Human - not judge it. Only a Human is supposed to judge another Human in person - this is live, fair, and humane.