Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Preparing for Malicious Uses of AI

blog.openai.com

Post reply on HN
Print view View original

151–160 of 233 posts

Re: Preparing for Malicious Uses of AI

#151

Post by notabee »

jcadam wrote: post #48

Evil startup idea gleaned from paper: Use AI/ML to scour a sales prospect's online persona (social media) and build a 'vulnerability profile' and generate targeted, personalized cold-emails (or even phone calls eventually). Also identifies 'levers' for a particular person that can be used to influence a buying decision. May even pre-qualify leads for you and tell you when not to waste your time :) I mean, a good sale…

Cambridge Analytica already did this for the past U.S. election. It had an app that told canvassers which doors were worth knocking on and what sort of dialogue to use depending on things like the person's estimated neuroticism and other traits. It'll only get more sophisticated.

Re: Preparing for Malicious Uses of AI

#152

Post by canaanical »

Diederich wrote: post #123
kypro wrote: post #35

Earlier quoted context omitted.

> I'm more concerned about Google using AI to mine every conversation I've ever had or my browsing history to classify me as a dissiden before I apply for a visa to travel to China or the United States, or as a deadbeat before I apply for a bank loan, or sick before I apply for insurance, or as unrehabilitatable before I apply for parole. I'm quite paranoid about this, yet whenever I speak to people about it either p…

> ... feel we've lost the battle for privacy. I agree with pretty much all of the concerns you and many others have cited, and I worry about that stuff too. On the other hand, I think about the past. How much privacy did most humans have in history? To be clear: I'm nothing like an expert on these topics. I'm specifically not considering the wandering nomad types. I think that the vast majority of all humans who have…

But the consolidation of energy (capital) and influence that allows corporate organisms to drive the "norm" has not been practiced on the scale of billions of users. "privacy" seems to be somewhat synonymous with, "finding a viable existence in parallel with the existing system." With massive-scale ability to automate who is within a norm or outside of it, avoiding an arbitrarily defined norm becomes un-viable.

Re: Preparing for Malicious Uses of AI

#153

Post by lowbloodsugar »

kypro wrote: post #35
Mizza wrote: post #30

The disturbing thing about this paper to me - flashy though it may be - is what they left out rather than what they kept in. OpenAI appears to only be thinking of the crimes-against-individuals segment of malicious AI, rather then the crimes-against-humanity type of malicious AI that the surveillance advertising corporations who are supporting OpenAI are building. I am far, far less worried about an assassin's drones…

> I'm more concerned about Google using AI to mine every conversation I've ever had or my browsing history to classify me as a dissiden before I apply for a visa to travel to China or the United States, or as a deadbeat before I apply for a bank loan, or sick before I apply for insurance, or as unrehabilitatable before I apply for parole. I'm quite paranoid about this, yet whenever I speak to people about it either p…

Well, the good news is that fossil fuels will run out soon and then we'll all be too busy dying of looting and starvation while old datacenters are used by local warlords as death mazes for entertainment.

Re: Preparing for Malicious Uses of AI

#154

Post by AndrewKemendo »

3pt14159 wrote: post #103

Reading the comments on this article make me realize how unserious most engineers and software developers are taking this. These threats are real. The weaponization of everything is actually happening. Since I wrote about it a month ago (Self-Crashing Cars) a number of people have reached out including people with actual insight into the military aspect of it. Militaries around the world are getting ready for true AI…

You can trust that people in the US Military are taking this very seriously. I know because I, alongside LtGen Shanahan and Eric Schmidt briefed Secretary Mattis personally about these issues and is part of my ongoing work within the IC/DoD.

You're correct though that most line DL engineers don't have these issues at top of mind. I don't know any ML researcher worth their salt that hasn't thought about them, but the tendency is to brush the concerns off until we're closer to more generalized systems.

That is not a wholly unreasonable position for many reasons, especially given the history of hype around AI, but I'd like to see more discussion happening between junior and mid-level engineers about these things - and especially more work being done about human-machine interfaces in the AI context, of which very little design thought put in.

Re: Preparing for Malicious Uses of AI

#155

Post by PoachedSausage »

shostack wrote: post #140
PoachedSausage wrote: post #130

Earlier quoted context omitted.

I recently noticed when coming back through UK passport/ID control that there are now boxes hanging from the ceiling in the waiting area, I assume to scoop up IMSI's/Wifi/Bluetooth/NFC MAC addresses from phones and tablets. All tied to passport/facial biometrics presumably.

Is simply turning off your device while there a solution? Asking as someone who may travel to the UK and EU this year.

I'm speculating, you could try asking GCHQ. With SDRs so compact and cheap these days the chances are high that they are installed in such places as airports and train stations. It is fairly safe to assume that turning things off should prevent most data leakage, but for the truly paranoid you would have to remove all the sewn RFID tags from your clothes and not carry contactless enabled cards.

Re: Preparing for Malicious Uses of AI

#157

Post by platz »

d0lph wrote: post #146

https://www.eff.org/files/2018/02/20/malicious_ai_report_fin... Still reading the paper and forming an opinion. But my initial thoughts are what exactly is new here that couldn't be done through some other means? I'm sure there will be interesting implications, but right now nothing seems particularly novel.

it is an injunction to policy makers, not a discovery of new methods.

Re: Preparing for Malicious Uses of AI

#158

Post by kakarot »

3pt14159 wrote: post #103

Reading the comments on this article make me realize how unserious most engineers and software developers are taking this. These threats are real. The weaponization of everything is actually happening. Since I wrote about it a month ago (Self-Crashing Cars) a number of people have reached out including people with actual insight into the military aspect of it. Militaries around the world are getting ready for true AI…

Here's the crux of the issue with respect to nuclear WMDs:

In order to prevent citizens and other countries from creating nuclear bombs, our government severely limits access to relevant tools and materials, and actively seeks to censor the knowledge of how to build modern nuclear devices from other parties.

In order to prevent citizens and other countries from creating dangerous rogue AIs, our government _____________

What do you think goes in that blank?

Re: Preparing for Malicious Uses of AI

#159

Post by platz »

landryraccoon wrote: post #148
nradov wrote: post #138

Earlier quoted context omitted.

Whatever. Let me know when someone invents an AI as smart as a lab mouse. Until then this is all just pointless idle speculation.

That's fair, but arguing something isn't possible is very different from arguing that when it happens you can solve it with a bomb. The conversation is about a hypothetical strong AI. You can say strong AI isn't possible, like you can say warp drives aren't possible, and it's still valid to discuss what it might be like if they were possible.

you're missing the point assuming we're talking about strong AI - when the paper is more addressing malevolent uses of weak AI.

Re: Preparing for Malicious Uses of AI

#160

Post by kakarot »

kypro wrote: post #35
Mizza wrote: post #30

The disturbing thing about this paper to me - flashy though it may be - is what they left out rather than what they kept in. OpenAI appears to only be thinking of the crimes-against-individuals segment of malicious AI, rather then the crimes-against-humanity type of malicious AI that the surveillance advertising corporations who are supporting OpenAI are building. I am far, far less worried about an assassin's drones…

> I'm more concerned about Google using AI to mine every conversation I've ever had or my browsing history to classify me as a dissiden before I apply for a visa to travel to China or the United States, or as a deadbeat before I apply for a bank loan, or sick before I apply for insurance, or as unrehabilitatable before I apply for parole. I'm quite paranoid about this, yet whenever I speak to people about it either p…

I will have lost my battle for privacy when I am six feet underground or in a straitjacket, and not a moment before. We can NOT accept defeat.
Post reply on HN