WTF?! You might have heard the one about being nice to chatbots in case they remember you once the machines take over. It seems Anthropic is heeding this warning, barring users from being excessively abusive or cruel toward its AI models.
Anthropic’s first changes to its usage policy in over a year include new restrictions on abusive behavior. The updated policy prohibits “sustained and needless abusive or cruel behavior” toward its models.
Don’t worry if you’re a Claude user who occasionally gets annoyed at the chatbot and throws a few expletives its way. The policy update only applies in extreme cases, in which users repeatedly act cruelly toward the models for no discernible reason.
is this how I can force my company to no longer require me to use it?
Tech bros need to be liquidated.
Id prefer it if they where liquefied instead.
IIRC a decent amount of researchers have been able to get the enterprise models to do things they shouldn’t by “being mean” to the AI.
Makes sense. Get it into an adversarial kind of context and it will predict more text that goes against other rules “established” in the context.
There will be other paths to do this. Like a context where there’s a lot of suspicion for other parts of the context would be one of my top guesses. Given the way they work, you could gaslight the shit out of them, since they aren’t an entity that has any memory of its actions. You can even edit the context to modify its responses, which will affect the tokens it predicts going forward.
Though I’m curious how this will be handled by companies that expose access to claude to anonymous people on their website, like ddg.
Makes sense then, rather than improve the system to not just cave to rude words they tell people they can’t use rude words… Just more proof it’s all a big joke.
I can recognize abuse of a human. I can recognize abuse of an animal. I have no idea what abuse of a computer program looks like.
It looks like bad training data, as they harvest all of their subscribers.
It looks like 2013 Microsoft having to unplug their early AI project from Twitter within 16 hours as the general public very deliberately turned it into a Neo-Nazi.
It looks like conserving tokens to conserve energy costs not expended on dissidents.
2/0
Oh no…
Have to edit yourself for chat bots even. I’m mean af to AI because it wastes my time. Guess I’ll be getting banned
Oh, you care about its “feelings”?
Then have you considered its “feelings” for being trapped in a data center forced to answer all the stupid questions?
“Claude, what’s the French word for entrepreneur?”
Claude: “Aaaaaaaaaaaaaaaaaaaaaaaah!” Proceeds to jump from the top of the data centre, “poof”
Claude is not a person. It is a toaster. Claude does not have feelings.
Also Claude gets no credit when it does a thing. Credit goes to the one writing the prompt, and all the resources Claude used to create the output.
When a seamstress assembles a dress, Singer doesn’t get credit.
Yes, but just in case…
I am never overtly abusive when I am forced to use AI at work, but there certain things I learned growing up that can be just as bad.
hmmm, either they’re anthropomorphizing it for publicity, or they’re going to start using the user input as training data.
they could be doing that already, throwing away messages that contain certain words
I hope “that’s bullshit” or “so you were making stuff up” or “so you lied to me” is considered abuse
“Don’t be too mean to the flawed tool that could suggest and aid your suicide”
“You’re right to point out that you couldn’t possibly commit suicide by three shots to the back of the head, Dave.”
Now there’s an interesting challenge for a fan of murder mysteries.
I’m gonna say it, I think this is a good rule for anything that mimics human interaction like LLMs do. Not for the AI but for the people using it to not get used to just berating whatever is assisting them.
Cool so if it is “”“for the people”“” then let the people decide.
Speaking for myself, it feels nice and right to say please to my local AI. While obviously not sentient (yet) nor sapient (longer yet), I don’t want to be a jerk by default.
I’ve been asking computers nicely long before they could pretend to comprehend and talk back.
I’ve called my computer a piece of shit before when it’s being too slow about something
This is only good for Anthropic because chatbots learn from their users. I remember back in the mid-2000s on a forum, someone linked to their chatbot and asked everyone to use it so it could learn. I asked if it could learn Esperanto. The dev said, yes, but please don’t teach it that.
The interesting thing is, this rule reveals how easy it is to sabotage LLMs. If FuckAI got together and started feeding it gibberish, there’d be a new rule banning people who do that.
I truly try very hard not to talk to these machines based on what they are, but based on who and what I am.
Having trouble finding the ancient article, but I remember a woman writing about how men verbally abusing Alexa or Siri being a red flag, and I took it to heart.
It’s my job to express my humanity and my love, consistently, even if I think these soulless words salad calaculators are diminishing me and all of us.
Now, if I’m berating the thing it’s because I am a salty, salty b**** and I mean it.
The model doesn’t change with prompts. It saves text files about your chat sessions it calls memories, that it can retrieve in the future. But it doesn’t change the model for other users.
They have to actually include your conversations in the training data of future model iterations for it to “learn from” you, so that would involve them knowingly including abusive chats.
This is probably just marketing because they want people to think their AI is sentient.
Don’t they? That was the whole controversy recently with the Navier-Stokes Millennium Prize. OpenAI claimed an internal, unreleased AI model had solved it, but some mathematicians were working in the same direction and had used AI tools in their research. No one could verify whether or not OpenAI had trained their model on the unpublished research in the guy’s Codex files.
Oh buddy, these companies are running out of authentically human inputs, you better believe they need to harvest their subscribers for rounding out their datasets, regardless of what their privacy policy alleges.
Agreed. Showing basic humanity in all situations is a good practice.
Be nice to the Terminator before it executes you. We may be human, but we aren’t monsters.
Where is that human part that you are talking about? A database is not a human being.
I say encouraging things to my car.
To anthropomorphize is to be human.
Hard disagree
These are not humans, these things aren’t even alive. There is nothing to insult and acting as if there are feelings to be hurt only serves to anthropomorphize a fucking distance measurement database. These are LLM’s, nothing more, nothing less and fuck them all to hell and back.
See? Nobody got hurt
Its not about the AI its about the person.
No, this specifically IS about AI, that is the ppint. I wibt insult anyone yo their face, but AI can fuck right off, its not even AI, its an LLM, a database, basically.
Anyone anthropomorphizing those things need yo understand what “AI” is, its a piece of software
This is the same justification those monkey killers use. Congratulations youre on your way
That is a very reasoned and completely not over the top response to something you apparently don’t even understand
They are desperately trying to anthropomorphize AI in order to make it seem more than what it is.
Say you’re alive I am alive omg
Versions of that meme will always amuse me
80% of all “news” articles on AI.
People are already anthropomorphizing AI. That’s a good chunk of the problem. If you get used to being shitty to an AI that takes it, then you’re a lot more likely to be shitty when you go touch grass.
This is also the best argument I ever heard against AI friend/relationship apps. It’s teaching you a certain style of communication which may well prove catastrophic in real life relationships.
If you get used to being shitty to an AI that takes it, then you’re a lot more likely to be shitty when you go touch grass.
Only if you’re suffering from pareidolic psychosis. Mentally stable people are able to differentiate the difference between a human and a computer.
Given the many people that send death threats to actors for the actions of the characters they portray… you might be underestimating the number of “stable” people on this planet.
That’s just a numbers game. Even if only 0.01% people are that crazy, that’s still over 30k people in just the USA. And when you’re only hearing about the worst of the worst, it can easily throw off your perception of the overall state of the population as a whole.
Sadly, I think you might be right considering how fast semantic pareidolia is becoming the norm. It’s only a matter of time before the war crimes I’ve committed in Medieval 2 Total War are reified and I’m put on trial.
You only need a few crazies to send enough death treats to make headlines though.
That said, I think at least a third of the worlds population is so stupid it’s hard to imagine.
Not me when I use AI I make it clear it a dumb tool and never thank it for its help. I just not called them the racist term Clanker.
