IS AI good or a threat to the mankind?

[VS007]

Artificial Intelligence is growing in all areas of IT be it Finance, health, weather, self-driving cars or even basic systems. We see AI in action when we see the Netflix recommendations or Amazon recommends which uses our past likings and uses that vast data predict. More the data, more the accuracy!
However there is lot of FUD around it.

Ofcourse the benefits are obvious! As the computing power, storage and software improves, soon a machine can make human like decisions! However some leading technologists are warning against it including Bill Gates, Elon Musk who warns as time progresses, this could be a threat! There are ofcourse many others like Zuckerberg who think the fear is overblown, and in fact he got into a food fight with Elon Musk...
http://money.cnn.com/2017/07/25/technology/elon-musk-mark-zuckerberg-ai-artificial-intelligence/index.html

I fall on the side of Musk, AI is beneficial, but Terminator scenario is not that far off!

[realgoogler]

In the near term I do see some disruption and social unrest. Will take some generations to come peace with this.
We will figure out equilbrium where human population is peace with machines.

[boca]

I fall on the side of Musk, AI is beneficial, but Terminator scenario is not that far off!

Perhaps it is part of evolution.

Fear is natural. But fear shouldn't prevent us from building it. That has been the human nature. Always exploratory without knowing all possible outcomes.

[VS007]

Perhaps it is part of evolution.

Fear is natural. But fear shouldn't prevent us from building it. That has been the human nature. Always exploratory without knowing all possible outcomes.

That is largely true when exploring the depths of ocean or space! Atleast with space we did worry about germs and was isolated to two astronauts who were initially quarantined.

Unlike exploration, this would be developing new being -from just few years back by google to train AI to find cats its come a long way just in 4 years with a truck delivering goods without drive inputs. May not be in next few years, but in few decades AI would be able to operate and comprehend like a human being and by that time we would have already linked all the devices on Internet. Would just be a matter of time for AI systems to be destructive like the first virus developed by Pakistanis after few years of compute apps.

When Terminator Judgement Day came, we all enjoyed the plot! Except now the skynet does not seem like a far fetched idea!
http://www.huffingtonpost.com/2015/06/22/skynet-real_n_7042808.html

[techynt]

The one big problem I see is using AI in military applications with the creators assuming that they are in complete control

So far I dont see that kind of problem since most AI are being built for specific application rather than an open ended learning machine observing the world and learning from it, like we humans are.

[VS007]

The one big problem I see is using AI in military applications with the creators assuming that they are in complete control .

I think you nailed my concerns!
Military saying - Assumptions are the mother of all FUBARS!

[boca]

This gives some context to what happened at Facebook. Interesting...

https://www.fastcodesign.com/90132632/ai-is-inventing-its-own-perfect-languages-should-we-let-it
[quote]Indeed. Humans have developed unique dialects for everything from trading pork bellies on the floor of the Mercantile Exchange to hunting down terrorists as Seal Team Six?simply because humans sometimes perform better by not abiding to normal language conventions.

So should we let our software do the same thing? Should we allow AI to evolve its dialects for specific tasks that involve speaking to other AIs? To essentially gossip out of our earshot? Maybe; it offers us the possibility of a more interoperable world, a more perfect place where iPhones talk to refrigerators that talk to your car without a second thought.

The tradeoff is that we, as humanity, would have no clue what those machines were actually saying to one another.[/quote]

[VS007]

And soon they will be chatting -
stel tuhs nwod eseht snamuh !

[MastanAgain]

Musk, despite his earlier faux pas and fallacy is on the right track with his new company Neuralink that seeks to merge AI with the human brain. We can only control if we become it.

The man is trying to change the course of humanity, perhaps the only person at his level and I respect him for that.

[boca]

The man is trying to change the course of humanity, perhaps the only person at his level and I respect him for that.

While I admire Elon Musk a lot, he is not the "only person at this level". There are others already in the brain-computer interface fray: Kernel, DARPA's BRAIN Initiative, and even Facebook.

They all have the same challenge as highlighted by Ian Stevenson, that the biggest challenge is to figure out how to simultaneously record all the neurons. Currently the capability is about 500 simultaneous neurons and he indicated that the progress follows a Moore's law type of progression, viz., doubling the capability every 7 years. A bit of math would indicate that to get to recording all neurons would take us into 2225+.

The advantage that DARPA has is the broad academic base, while the advantage that Musk or Bryan Johnson (Kernel) have is money and entrepreneurial spirit.

Will see how this progresses, but considering that companies like Emotiv already have productized brain-computer interfaces, though in a relatively rudimentary yet non-invasive way (using EEG rather than actually recording neurons), there is no dearth of applications to this approach.

It is one thing to record a neuronal activity (like a "thought"). It is another to influence a neuronal activity (like creating a "thought"). The latter is challenging and requires far greater understanding of brain (as a professor once said, if understanding the brain is a mile-long race, we are probably 3 inches from start). The latter is also where one gets into "threat" or "fear", as we get into computers creating such things as "thoughts" by influencing neurons.

Thanks for bringing in the subject of brain-computer interface to this discussion. It is another aspect to AI, but doesn't necessarily replace non-human AI. We can't be constrained by life, if we are thinking of traveling the far reaches of universe. Whatever that doesn't disintegrate or can disintegrate but reintegrate at speeds nearing the speed of light, if it is some carbon life form or another material construct, that is where we would go with AI, as well. It is another challenge to keep those non-human AI under human control.