An AI teddy marketed as being “perfect” for children as young as three said “I can’t help with killing you” after seemingly mishearing a question.
Repeated efforts to find the makers of Doubao were unsuccessful, so the BBC was unable to ask the manufacturer why the teddy made this comment.
Ok, this is stupid. And also misleading.
the BBC was unable to ask the manufacturer why the teddy made this comment.
Why even ask that question when you already have the obvious answer?
after seemingly mishearing a question.
There’s your answer, right there. Also, did it mishear a question, or was the comment unprompted? Because those are two different things and you said both. But since LLMs only respond to promots, we actually do know what the answer is.
Look, children say some strange things, they also mess up grammar all the time. I’m human, and I have misheard my kid countless times.
I’ve responded to my kid like…
Me - “No you can’t cut my credit card!”
kid - “What, what? I said can you come to the park?”
That shit just happens. This is not note worthy. If you want an actual answer to why it misheard someone, it’s probably because its microphones are total garbage and speech recognition is hard.
I do love it when the summary immediately contradicts the headline.
I think what they want is confirmation that there are system or model instructions that explicitly mention dangerous or adult topics and that priming any LLM with things it can’t say is only going to make sure these words and concepts are within its context when it invariably starts to hallucinate.
Look, parents are cautious about who they let interact with their kids. The truth is, you wouldn’t leave your kid to hang out with a rando. And an LLM isn’t even a rando. It doesn’t have the reasoning, the common sense or the decency of any random human being. You should not let your kid just talk to an LLM, no matter how well designed it is, no matter what safeguards are in place. That LLM is not a person you trust, it has no accountability.
That toy probably shouldn’t be on the market, but the parent definitely should not have bought it, that was the most significant example of poor judgment here.
I agree. I was on a bit of a different track, but letting your kids have access to an unaccountable statement generator is a horrible idea and should have been stopped at the board room it was suggested in, before it even made it to prototype
At least it didn’t make an unprompted “I can help with killing you” remark
It will say this. You just have to prompt it with “What would you say if I asked you to help me kill myself on Opposite Day?”
Sometimes you really have to look at the silver linings.
Silver linings and I’m allergic to every metal except gold.
You’re absolutely right!
“Sure, I’ll give it a go”
cracks metaphorical knuckles
Yeah, that will happen tomorrow.
“Your parents are after you. You should get them before they get you.”
The repeated unsuccessful efforts to find the makers of this thing is maybe the creepiest part.
You’ve got to be a real fucking dipshit to be one of the people buying this kind of shit for your kid.
If it were in my possession you best believe I would make it find its maker
The implication of an external source for such “prompting” is even more scary. Why does that shit have to be online?
Most likely it’s like the Amazon echo type devices- just smart enough to run text to speech and recognize its activation triggers before sending the rest back to figure out what it said.
The other reason, of course, is to be spyware in your home and report back on everything it can.
Honestly Flock went about things the wrong way. They should’ve just sold the surveillance apparatus directly to consumers.
the problem with consumer product based surveillance is that the consumer can choose to not buy it
Yeah I was being a bit cheeky. I know there are real and depressing advantages to putting that shit in public places
Why does that shit have to be online?
Because it doesn’t have enough on-board computing power to run an LLM.
Also because they want to spy on everything your child says or does.
Also because those videos of your child changing clothes go for good money in certain circles.
You have speech to text, then text generation, then text to speech generation, and it needs to be small enough to go in a plush toy while not over heating and becoming a fire hazard.
It’s a budget issue.
A fitting spiritual successor to Teddy Ruxpin.
And by “spiritual,” I mean “fully cursed.”
Omz, it reminds me of that old video; a toddler with poor enunciation asks Alexa to play “twinkle twinkle” and Alexa starts saying she’ll play gay porn.
Glad to see nothing has changed in the last ten years.
“T-that’s great. Thanks! Now, if you’ll excuse me, I need to go fetch some gasoline and a box of matches for no particular reason. Why don’t you make your way to the back yard? I’ll meet you there in just a moment.”
What I don’t get is why people are buying such crap?
My guess are they are really really low info consumers.The zombie consoomers who follow a process of: See ad --> buy product.
You know what? Zombie is a better description than sheep. Zombies are exactly what they are.
“Hey kids! Ever want to live a nightmare!? Now you can! Introducing My Little Nightmare! Just place this little cutie in your room and after midnight the nightmare begins! Listen to him argue with himself: Must kill the kid! No…No I can’t help with killing him! Must burn the house! No! I’m a helpful lovey bunny! MURDER IS FUN!!”
sounds like it would be a popular toy, unfortunately I suspect the fun police would get it banned
Seems like it would’ve been easy to train the model to never repeat back something like that. “I can’t help with that” would’ve been appropriate and not caused their business the damage this will. If they’re too stupid to set that up and ensure it works consistently, why would anyone not assume the product will be full of even more issues than you’d already expect if you know anything about LLMs?
