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

The NSA’s machine learning algorithm may be killing thousands of innocent people

arstechnica.co.uk

Post reply on HN
Print view View original

81–90 of 221 posts

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#81

Post by afsina »

graycat wrote: post #74
afsina wrote: post #72

Oh come on "machine learning algorithm may be killing thousands.." just put a "may be" and your BS assertion becomes more plausible? By looking at this, people are having and idea that attack drones (it is rich that article puts one picture of it in the beginning) are loaded with such software and killing by looking at the result of a classifier. Or some super computer gives you a name and says "exterminate". Apparen…

> If it is a false positive, probably they just ignore it. And how do they know it's a false positive? Indeed, if they really knew that, then they would have some more training data.

I am not defending peoples lousy/callous decision processes for fatal actions. But telling that bad machine learning algorithms are responsible for killings is just stupid.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#82

Post by random_upvoter »

Houshalter wrote: post #7

They literally named it Skynet. They have an evil sense of humor. Actually using machine learning to detect terrorists isn't a terrible idea. But you are going to get an error rate, and probably a high one in the noisy real world. Maybe only 50% of the people you detect are actually terrorists. Maybe it's even worse than that. We can't even test it because there is no validation set and unreliable labels. The reasona…

Reading discussions on drone strikes makes you wonder what ever happened to pacifism. This whole idea that you can just go around other countries killing people -- even if they intend on killing you, and even if you would be able to hit the right target 100% of the time-- it's just old. It's primitive. And that is not to blame the USA. At this point in history humanity, as a collective, has the maturity of an 11 year old. Maybe it's time to evolve a bit.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#83

Post by MikeNomad »

oafitupa wrote: post #67
Aeolos wrote: post #58

Earlier quoted context omitted.

Exactly. In which freaking world are assassinations without due process legal or even desirable? The mere fact that we are debating whether someone was or wasn't assassinated is a perversion in itself. Arrest them and try them in a court of justice. End of story.

Arrest them? They were in freaking Pakistan. Why would you arrest them? You guys are going crazy. Don't be surprised when you travel and realize the world hates you.

I already know parts of the world hate me, because I come from the country that engages in drone strikes.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#84

Post by akerro »

stevetrewick wrote: post #73
akerro wrote: post #63

Earlier quoted context omitted.

- political dissidents https://en.wikipedia.org/wiki/McCarthyism And that's what it is used for. Australian gov. used their abilities to mark people who say against coal energy as anti-government, all they did was support renewable sources of energy. https://overland.org.au/2014/07/surveillance-of-activists-is... and Canada: http://www.theguardian.com/environment/2013/feb/14/canada-en... - minority groups FBI was spy…

>Australian gov. used their abilities to mark people who say against coal energy as anti-government Even the partisan sources you've linked don't support that particular assertion AFAICS. It appears that gov.au placed a group of activists with the potential intent to disrupt energy supplies under surveillance. 'Green' activists suffer from a bad case of noble cause corruption which prevents them from understanding th…

I don't know any good sources about it. I heard about it somewhere on reddit probably and it stuck in my head, there are more different kind of sources if you need more:

http://www.theguardian.com/environment/earth-insight/2014/ja...

http://www.abc.net.au/news/2012-01-07/brown-slams-spying-on-...

http://www.smh.com.au/environment/afp-spies-targeting-green-...

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#85

Post by afsina »

vladtaltos wrote: post #78
afsina wrote: post #72

Oh come on "machine learning algorithm may be killing thousands.." just put a "may be" and your BS assertion becomes more plausible? By looking at this, people are having and idea that attack drones (it is rich that article puts one picture of it in the beginning) are loaded with such software and killing by looking at the result of a classifier. Or some super computer gives you a name and says "exterminate". Apparen…

the problem is you don't know that 'If it is a false positive, probably they just ignore it.' -> 'probably' is a fucking optimistic view. AND how the hell do you know this is a false positive ? some algorithm tells a soldier that this person is probably a terrorist - what happens then ? do you think they bother to verify this by spending more money ? In my opinion it is most likely that, the people using this softwar…

AFAIK for such decisions human information/intelligence is required. And there are often screw ups/bad decisions. No need for bad machine learning algorithms for this.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#87

Post by marcosdumay »

Houshalter wrote: post #7

They literally named it Skynet. They have an evil sense of humor. Actually using machine learning to detect terrorists isn't a terrible idea. But you are going to get an error rate, and probably a high one in the noisy real world. Maybe only 50% of the people you detect are actually terrorists. Maybe it's even worse than that. We can't even test it because there is no validation set and unreliable labels. The reasona…

How many terrorists are there to put on the List?

For something around a ten of thousand people, at a minimum 80% of them will be innocent, just by sheer lack of actual terrorists to fill that many slots. And that's assuming a perfect fitting algorithm, with absolutely no false positives.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#88

Post by mike_hearn »

nl wrote: post #53
Houshalter wrote: post #7

They literally named it Skynet. They have an evil sense of humor. Actually using machine learning to detect terrorists isn't a terrible idea. But you are going to get an error rate, and probably a high one in the noisy real world. Maybe only 50% of the people you detect are actually terrorists. Maybe it's even worse than that. We can't even test it because there is no validation set and unreliable labels. The reasona…

The reasonable thing to do with that information, would be to surveil them further, search their house, or arrest them. Not assassinate them without a trial. This appears to be what they do? The National Security Agency reportedly tracked phone calls between the courier Abu Ahmed al-Kuwaiti's relatives in the Persian Gulf to all numbers in Pakistan, and NSA surveillance eventually tracked Abu Ahmed al-Kuwaiti's locat…

The evidence is that the head of the NSA literally said, "we use metadata to kill people". It doesn't get more direct than that. Signature strikes are a well documented phenomenon by now.

Also there's no way the CIA drone program can sustain the kind of long term kill rates they've been having without routinely killing innocent people. Terrorists just aren't that common (although if you define anyone who tries to seek revenge thanks to an errant drone strike as a terrorist, that "problem" can eventually solve itself).

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#89

Post by vlehto »

agd wrote: post #13

Big data analysis + mass surveillance is a frightening prospect. Of course you can train software to look for 'terrorists', but you could also train it to look for: - whistleblowers - minority groups (e.g. gay people, particular religious beliefs, political affiliation) - political dissidents - journalists whose behaviour changes - personal vulnerabilities (affairs, mental health issues etc) Think what authoritarian…

Big data analysis + mass surveillance is just more efficient than previous methods.

Let's imagine: 1. citizens have right to fair trial 2. terrorist hunting is not done by police, but by military intelligence organization that cannot arrest, attack or sue anybody, just gather data and inform police/military or prosecutor 3. that military organization is supervized by democratically elected guy (president) 4. Law that defines clearly when such information can be used, and defensive attorneys are not excluded from using it.

That organization being part of military is inportant. If it's actually clandestine organization with priority of defence, it should be unwilling to reveal it's capability to foreign players. Unless it's actually matter of national security. This is probably happening already with multitude of military forces and we never hear about it. Shit only starts when police starts doing same stuff while disregarding fair trial. SS-troops and KGB we're pretty good at fucking up people even prior to mass surveillance and big data analysis.

Re: The NSA’s machine learning algorithm may be killing thousands of innocent people

#90

Post by donatj »

We're killing people in a country were not even at war with over statistics. Not evidence, but statistics. People who statistical might have something against us. Since when has this been grounds for killing someone? This is utterly terrifying.

Just imagine if someone decided to do this to us. How does Pakistan feel about all this?

This is approaching Auschwitz levels of evil.

Post reply on HN