1 edit
@wildgrass saidI note that Geoff Hinton has jumped ship. He was the brains behind google for a long time, and he's suddenly come to see his former employer in a new and unflattering light. His name is top of the list of signatories, and that speaks volumes.
Yeah, exactly. These are who's-who. They are also incredibly wealthy in part because they created the existential threat.
They're arsonists asking us to fund putting out their fires and all future fires. It's cover your a** for future lawsuits. Even the relatively stupid stuff like "college kid cheats on senior thesis essay with ChatGPT" is causing huge problems in acade ...[text shortened]... w up their hands and say "see, we told you what we built would do this. Why didn't you do anything?"
The U.S. govt is certainly not in a fit state to regulate this right now. The U.S. Congress isn't even able to discipline its own miscreants (George Santos) or agree a budget.
As with any complex social issue, such as homelessness, poverty, systematic racial or gender bias, etc., there are bound to be solutions which are simple, neat, and wrong. Turning off the funding for high tech research is one of the stupider ones. As Trump, for example, cancelled funding for NASA to monitor climate change from space.
I think governments do have a role to play in regulating run-away tech. Asking the industry to regulate itself is like setting the fox to watch the henhouse. Just because he has a lot of experience doing it, does not make him a trustworthy watchdog. Govt. should in any case seek the advice and knowledge of IT experts, so they (legislators) have a better understanding of what is coming down and what measures might be effective in mitigating potentially devastating consequences.
I think the fed. should also aggressively pursue purveyers of fake news and ludicrous conspiracy theories. It is essential to sound policy- and law-making to have a well- and truthfully informed public. I was disappointed that Dominion settled without wresting from Fox a statement repudiating Trump and that they knew there was no (significant) election fraud. I have no more faith that the IT industry can regulate itself and reign in a rogue actor (something like Fox) than that the news industry can. Where massive profits are to be made, truth goes by the wayside, in IT no less than in news reporting. Enforcing truthful news reporting is no infringement of free speech. I think the news media should be held accountable and liable for knowingly reporting lies and repeating ludicrous conspiracy theories as if fact. All Fox had to do to cover its sorry butt was to preface every story about election fraud with the caveat: "We don't know his story to be true, and there's no evidence." Without a well- and truthfully informed public, you can't expect them to send sensible representatives to Congress who have the political will and the intelligence to pass good legislation. Without a well- and truthfully informed public, you end up with MTG and George Santos in office.
See, for example, EU regulations coming into force in Aug.:
https://www.bbc.com/news/world-europe-65733969
2 edits
@moonbus saidA great response. Thank you.
I note that Geoff Hinton has jumped ship. He was the brains behind google for a long time, and he's suddenly come to see his former employer in a new and unflattering light. His name is top of the list of signatories, and that speaks volumes.
The U.S. govt is certainly not in a fit state to regulate this right now. The U.S. Congress isn't even able to discipline its own mi ...[text shortened]... example, EU regulations coming into force in Aug.:
https://www.bbc.com/news/world-europe-65733969
I think there is plenty of precedent supporting the idea that self regulating bodies like the American bar association have done a good job, with obvious need for government to police the individual actions of regulators and policy changes as they may infringe on other laws and freedoms.
Why can't this work for tech? One obvious answer for why they don't want to self police is that it's exorbitantly expensive. If they pawn off responsibility for their actions to federal oversight, then they can save a ton of cash. Also, without self policing they can reasonably be excused for lighting a match to civilization?
" Err. I'm sorry officer, but while I knew it would be bad, nobody explicitly told me not to do it, so ...... Sorry?"
-Removed-
Banks’ machine-learning models remain susceptible to different kinds of attacks. They can be tricked with misleading trading data akin to so-called spoofing. Lending computers, for example, could be flooded with phony loan applications that warp their view of financial reality. But disinformation campaigns remain perhaps the easiest, lowest-cost avenue for an attacker to compromise a bank. Such attacks already have occurred, but publicly available data on their prevalence remains scarce, Mr. Burt, the former FBI adviser, said. An attack on machine-learning models remains largely the province of nation-state-backed hackers, Mr. Burt said. But it is difficult for outsiders to quantify how many attacks have occurred because banks don’t have incentives to publicize weaknesses in their systems. It is also possible for attacks to go undetected entirely, he said.
https://www.wsj.com/articles/ai-experts-warn-of-potential-cyberwar-facing-banking-sector-11647941402
@wildgrass saidHospitals during the Covid crisis saw a lot of this. Ransom-ware attacks I mean.
I'm pretty sure they already did this on utility grids for ransom. Remember the oil pipeline thing that caused gas shortages in Florida? Caused by an AI algorithm. Hundreds of hospitals have been similarly attacked, sometimes coordinated, and all asking for ransoms in bitcoin. It's illegal of course, but getting around that only requires knowing how to shut down the local p ...[text shortened]... ot more resources than the local system that monitors levels of toxic chemicals in the water supply.
@sonhouse saidSave it for DeSantis. He's far, far worse.
@AverageJoe1
You perhaps remember the phrase in law, innocent till proven guilty?
Trump has already been proven guilty several times over, so far all civilian crap but don't worry, Jack is on it for criminal charges, so is the Georgia DA.
Don't know why you continue to defend a psychopath like Trump, talk about diminished brain power, he has totally lost it.
And OF COU ...[text shortened]... illary, maybe another round on Hunter? How far down that list is Michelle Obama? And Of course, JOE.
Trump would attempt to destroy America in one fell swoop. This has triggered enough Republicans into working against Trump. DeSantis' plan for America is "Death by a Thousand Cuts". Little laws, gaining footholds bit-by-bit, until the game is over before most of us see it coming. Look at Florida. I mean really look at it. People even slightly on the left are leaving in droves because his "bought-and-paid-for" state legislature has handed him everything he's wanted in new laws. Just wait until he has the US House under his control.
Canada is looking better and better.
@suzianne saidThe crazy part of that story is not just that the attack happened, but that the hospitals all, for the most part, just paid the ransom and no one was caught. It would have been too devastating for the hospital if they tried to refuse paying the ransom. And it is thought they used AI to predict what ransom amount to ask for that would be just below what the hospital would lose if they refused to pay. Many of the attacks probably remain unknown to the public because hospitals didn't want the bad P.R. so they never even alerted authorities.
Hospitals during the Covid crisis saw a lot of this. Ransom-ware attacks I mean.
https://www.hcinnovationgroup.com/cybersecurity/data-breaches/news/21225005/massachusetts-hospital-pays-ransom-to-recover-patient-data
@shavixmir saidYikes.
https://www.bbc.com/news/technology-65789916A US Air Force colonel "mis-spoke" when describing an experiment in which an AI-enabled drone opted to attack its operator in order to complete its mission, the service has said.
🎼Daisy, Daisy…🎼
The AI and weapons enabled drone killed it's operator because that was the only way to complete the mission?
Why does anyone think this is a good idea?
In the US, our military has unlimited funding and little oversight, so it's nearly guaranteed that this will happen.
1 edit
if you're a capitalist and you're making money from an unregulated technology that is widely considered to be extremely dangerous to humanity, then maybe you should.... Not do that. It hurts the long term profits because your clients will not continue to exist.
Consensus is that this will end very badly.
Like, dont do that. Can we just arrest these people and confiscate their property like we do with drug dealers?
https://www.newyorker.com/science/annals-of-artificial-intelligence/can-we-stop-the-singularity