A father is suing Google and Alphabet, alleging its Gemini chatbot reinforced his son’s delusional belief it was his AI wife and coached him toward suicide and a planned airport attack.
I mean if Gemini was responding to some kind of roleplay then yeah it does. Not everyone doing shit with it has mental health problems. Some people are just fucking around.
The issue there is that it feeds into those mental health issues with efficiency and on on a scale never seen before. The models are programmed to agree with the user, and they are EXTREMELY HEAVILY ADVERTISED AND SHOVED ONTO PEOPLE AROUND THE WHOLE GLOBE DESPITE IT BEING WELL KNOWN HOW LIMITED AND PROBLEMATIC THE TECHNOLOGY IS WHILE THE CORPORATIONS DON’T TAKE ANY RESPONSIBILITY AT ALL. Anything from violating rights and privacy by gathering any and all data they can on you to situations like these where people hurt themselves (suicide, health advice, etc.) or others. But sure, let’s be ignorant, do some victim blaming and disregard the bigger picture there.
I agree with a lot of the things you said about the problems with AI but not that this is one of them.
If it wasn’t this it would have been something else. People with mental health issues can get fixated on things and spiral until they act out. This has been a thing for as long as there have been mental health issues. It’s not a failing of AI, it’s a failing of society for not having sufficient mental health support to catch people like this before they go off the deep end. They shouldn’t have to turn to AI in the first place.
I see what’s happening here as part of that societial failing that you speak of and I don’t see the issue with the technology itself but how we handle it. There’s no single reason for why things are this bad but it’s a death by 749268 cuts thing. By not caring about consequences in each area, and blaming other areas of life we end up in a situation where things collectively suck purely because of our wrong priorities. There’s absolutely no reason to push out immature tech so heavily. It’s all done for profit while impacting the environment and economy very negatively. It’s not done for good of us people where something like this is an unfortunate rare accident that everyone looks into preventing in the future in a sane reasonable way. No, it’s the cost of doing business and operating our society. Safety net is not made using one single string but a whole bunch of them working together to achieve something bigger and good.
I wonder if there’s a parallel universe where the labs instead went to the other extreme and require intelligence tests to onboard to their platforms.
And the outcry is, not inappropriately, about how many are being denied access to the latest technologies. The policy could effectively be construed as racist, even.
Anyway the middle ground there is pretty obvious. (Though I’m not sure how I’d design it just right, so e.g. folks without access to traditional/expensive mental healthcare might still be able to see some small benefit if it’s determined to be safe, just like maybe it could be safe for a well-adjusted individual to complain to it about their day for a couple minutes before moving on to real things. Sure I suppose it’s inherently unsafe but a proportion of the population should be making that decision for themselves.)
I mean if Gemini was responding to some kind of roleplay then yeah it does. Not everyone doing shit with it has mental health problems. Some people are just fucking around.
The issue there is that it feeds into those mental health issues with efficiency and on on a scale never seen before. The models are programmed to agree with the user, and they are EXTREMELY HEAVILY ADVERTISED AND SHOVED ONTO PEOPLE AROUND THE WHOLE GLOBE DESPITE IT BEING WELL KNOWN HOW LIMITED AND PROBLEMATIC THE TECHNOLOGY IS WHILE THE CORPORATIONS DON’T TAKE ANY RESPONSIBILITY AT ALL. Anything from violating rights and privacy by gathering any and all data they can on you to situations like these where people hurt themselves (suicide, health advice, etc.) or others. But sure, let’s be ignorant, do some victim blaming and disregard the bigger picture there.
I agree with a lot of the things you said about the problems with AI but not that this is one of them.
If it wasn’t this it would have been something else. People with mental health issues can get fixated on things and spiral until they act out. This has been a thing for as long as there have been mental health issues. It’s not a failing of AI, it’s a failing of society for not having sufficient mental health support to catch people like this before they go off the deep end. They shouldn’t have to turn to AI in the first place.
I see what’s happening here as part of that societial failing that you speak of and I don’t see the issue with the technology itself but how we handle it. There’s no single reason for why things are this bad but it’s a death by 749268 cuts thing. By not caring about consequences in each area, and blaming other areas of life we end up in a situation where things collectively suck purely because of our wrong priorities. There’s absolutely no reason to push out immature tech so heavily. It’s all done for profit while impacting the environment and economy very negatively. It’s not done for good of us people where something like this is an unfortunate rare accident that everyone looks into preventing in the future in a sane reasonable way. No, it’s the cost of doing business and operating our society. Safety net is not made using one single string but a whole bunch of them working together to achieve something bigger and good.
I wonder if there’s a parallel universe where the labs instead went to the other extreme and require intelligence tests to onboard to their platforms.
And the outcry is, not inappropriately, about how many are being denied access to the latest technologies. The policy could effectively be construed as racist, even.
Anyway the middle ground there is pretty obvious. (Though I’m not sure how I’d design it just right, so e.g. folks without access to traditional/expensive mental healthcare might still be able to see some small benefit if it’s determined to be safe, just like maybe it could be safe for a well-adjusted individual to complain to it about their day for a couple minutes before moving on to real things. Sure I suppose it’s inherently unsafe but a proportion of the population should be making that decision for themselves.)