4 ms·

Artificial Intelligence Isn’t a Threat – Yet

9 points by softdev12 12y ago

nitin_flanker 12y ago
This is creating so much buzz.. I would like to say that just wait and watch. Time will tell whether it is or whether it is not
blazespin 12y ago
In one respect, AI is already damaging the sense of self worth of swathes of humanity by making them economically redundant. This will be an accelerating threat for sometime.
- Animats 12y ago
  That already happened, 140 years ago, with John Henry and the steam drill. https://en.wikipedia.org/wiki/John_Henry_%28folklore%29 https://en.wikipedia.org/wiki/John_Henry_%28folklore%29
- Bluestrike2 12y ago
  While I can empathize with the emotions behind that argument, it's rather hard to be persuaded by it. Jobs can become redundant, but people can't insofar as they can be retrained. If we're talking about what harms people's self worth, I'd imagine that telling people they're redundant is a lot more dehumanizing than having them retrain for new positions. In any case, human beings are remarkable adaptive. It's one of our most impressive traits; don't sell it short.
  - zyx321 12y ago
    Easier said than done. What are they going to retrain to? As AI continues to advance, more and more jobs will necome redundant. CGP Grey did a great video on this[1]. Even if you can't completely replace humans, often times, obe person will be able to do the job that used to take several. Where do those others go? What will we as a society do if for example AI-based Expert Systems make 90% of accountants and office clerks redundant? We can't retrain all of them to become doctors and scientists, can we? [1] http://youtu.be/7Pq-S557XQU http://youtu.be/7Pq-S557XQU
Animats 12y ago
The Wall Street Journal will consider AI a threat when an AI running a hedge fund starts firing CEOs.
hurtmyknee 12y ago
There are seven questions in this article, and topics like machine learning and program verification are in quotes. I don't think a competent engineer was consulted on this article. Perhaps as a community we should elucidate in a more formal fashion to the press.
GuiA 12y ago
New rule: if you have a PhD in "artificial intelligence" (aka mostly machine learning these days, although they like to say "artificial intelligence" as it sounds more futuristic to the layman) and you give interviews/write articles about the singularity/sentient AI/any of that other baloney, you lose your PhD. This is just mental masturbation for nerds, on par with drivel like Roko's basilisk (an example of the exact same logic being applied but to comedic ends is MIT's time traveler convention). Anyone who knows how to program knows that statements like: "Until we have some mechanism for guaranteeing that machines never try to replace us, or relegate us to zoos, we should take the problem of AI risk seriously." are just complete sensationalist bullcrap that have a chance of happening in the next century of exactly 0%. Many questions, mostly legal, come with things like driverless cars. But in all cases, those questions concern how human responsibility should be handled, not the risk of Battlestar Galactica becoming reality.
ankurdhama 12y ago
People don't really understand AI. I wrote a post about why AI is not a threat and the arguments are more of logical and fact driven then philosophical. https://medium.com/@ankur_dhama/artificial-intelligence-a-threat-d525799f912b https://medium.com/@ankur_dhama/artificial-intelligence-a-th...
ankurdhama 12y ago
People don't really understand AI. I wrote a post about why AI is not a threat and the arguments are more of logical and fact driven then philosophical. https://medium.com/@ankur_dhama/artificial-intelligence-a-threat-d525799f912b https://medium.com/@ankur_dhama/artificial-intelligence-a-th...
vixen99 12y ago
'AI will not be a separate intelligent entity, it will be the extension/evolution of existing intelligent entity, that is us.' With respect to intellectual threat I don't think you can say that. Your assumption is that any behaviour resulting from the creation of a separate intelligent entity (no one has the slightest idea how to do this) will be predictable (or be constrained by us) - having evolved 'from us'. Why should that be?
legohead 12y ago
I've been giving this a lot of thought lately because of Musk and Hawking. I think they are wrong, for a couple simple reasons. 1) A true AI wont care about us, or probably anything. It may just suicide. 2) Very simple to destroy. EMP, etc.

9 points by softdev12 12y ago

This is creating so much buzz.. I would like to say that just wait and watch. Time will tell whether it is or whether it is not

blazespin 12y ago

In one respect, AI is already damaging the sense of self worth of swathes of humanity by making them economically redundant. This will be an accelerating threat for sometime.

Animats 12y ago
That already happened, 140 years ago, with John Henry and the steam drill. https://en.wikipedia.org/wiki/John_Henry_%28folklore%29 https://en.wikipedia.org/wiki/John_Henry_%28folklore%29
Bluestrike2 12y ago
While I can empathize with the emotions behind that argument, it's rather hard to be persuaded by it. Jobs can become redundant, but people can't insofar as they can be retrained. If we're talking about what harms people's self worth, I'd imagine that telling people they're redundant is a lot more dehumanizing than having them retrain for new positions. In any case, human beings are remarkable adaptive. It's one of our most impressive traits; don't sell it short.
- zyx321 12y ago
  Easier said than done. What are they going to retrain to? As AI continues to advance, more and more jobs will necome redundant. CGP Grey did a great video on this[1]. Even if you can't completely replace humans, often times, obe person will be able to do the job that used to take several. Where do those others go? What will we as a society do if for example AI-based Expert Systems make 90% of accountants and office clerks redundant? We can't retrain all of them to become doctors and scientists, can we? [1] http://youtu.be/7Pq-S557XQU http://youtu.be/7Pq-S557XQU

Animats 12y ago

The Wall Street Journal will consider AI a threat when an AI running a hedge fund starts firing CEOs.

hurtmyknee 12y ago

There are seven questions in this article, and topics like machine learning and program verification are in quotes. I don't think a competent engineer was consulted on this article. Perhaps as a community we should elucidate in a more formal fashion to the press.

GuiA 12y ago

New rule: if you have a PhD in "artificial intelligence" (aka mostly machine learning these days, although they like to say "artificial intelligence" as it sounds more futuristic to the layman) and you give interviews/write articles about the singularity/sentient AI/any of that other baloney, you lose your PhD. This is just mental masturbation for nerds, on par with drivel like Roko's basilisk (an example of the exact same logic being applied but to comedic ends is MIT's time traveler convention). Anyone who knows how to program knows that statements like: "Until we have some mechanism for guaranteeing that machines never try to replace us, or relegate us to zoos, we should take the problem of AI risk seriously." are just complete sensationalist bullcrap that have a chance of happening in the next century of exactly 0%. Many questions, mostly legal, come with things like driverless cars. But in all cases, those questions concern how human responsibility should be handled, not the risk of Battlestar Galactica becoming reality.

ankurdhama 12y ago

People don't really understand AI. I wrote a post about why AI is not a threat and the arguments are more of logical and fact driven then philosophical. https://medium.com/@ankur_dhama/artificial-intelligence-a-threat-d525799f912b https://medium.com/@ankur_dhama/artificial-intelligence-a-th...

ankurdhama 12y ago

vixen99 12y ago

'AI will not be a separate intelligent entity, it will be the extension/evolution of existing intelligent entity, that is us.' With respect to intellectual threat I don't think you can say that. Your assumption is that any behaviour resulting from the creation of a separate intelligent entity (no one has the slightest idea how to do this) will be predictable (or be constrained by us) - having evolved 'from us'. Why should that be?

legohead 12y ago

I've been giving this a lot of thought lately because of Musk and Hawking. I think they are wrong, for a couple simple reasons. 1) A true AI wont care about us, or probably anything. It may just suicide. 2) Very simple to destroy. EMP, etc.