Whether it’s AI or some other technology that is projected to be highly integrated into our lives, should have laws and of course, transparency. It’s hard not to trust the tech you’re using, right?

 

Thats true, unfortunately it is in a lot of cases handled by the wrong people, in the wrong way and no one can say or do anything.

This is so very true - sad but true. You would think we had learnt from all those Hollywood movies! But apparently not

 

Whether it’s AI or some other technology that is projected to be highly integrated into our lives, should have laws and of course, transparency. It’s hard not to trust the tech you’re using, right?

Not just that. I think rules are a must for its own progress too. You can't make it perfect without guidelines.

 

Not just that. I think rules are a must for its own progress too. You can't make it perfect without guidelines.

Sure but rules are made to be broken. People and hopefully future AI machines will normally obey the rules that are in their own (sometimes misguided) self interest at a point in time. Rules/guidelines that are unreasonable for the situation will soon fail.

This is a example: https://forum.allaboutcircuits.com/threads/thought-for-the-day.44743/page-28#post-1312166

 

Rules/guidelines that are unreasonable for the situation will soon fail.

So you think, AI has no scope in future? If we go by this philosophy, there must be no industry standards.

 

I'm saying it's dubious that scientists and engineers will design systems that are angels on all respects with an absolute respect to rules. Government (and other criminal element) systems WILL be weaponized, and their A.I. WILL have an almost impregnable “off switch.” They WILL break in, steal, commit 'crimes' for the 'greater good' and have the power to subvert other AI systems in the pursuit of its goals.

 

Codifying any laws ... and the question is who makes them ... they are still implemented by the weakest link .... the human. Humans still falter to the various vices.

 

Anyone remember a sci-fi movie called "The Forbin Project" (Iu think thats right). Postulates a super computer that is given control of all nukes in the US. Ends up deciding that humans are not capable of making correct decisions. Gets its way by threatening to launch nukes on selected US cities if not obeyed. Think about it!. If it had( or could force) suitable sensors to be installed in every silo,could force its humans slave to ensure it had continuous power etc, there would be no way to shut it down without causing a nuke holocaust.In the movie , it has a general executed trying to shut to it down. Ends up forcing humans to connect it to a similar Russian super computer and they take complete charge.
AI gone overboard. Can't happen you say!. How would you prevent it!

 

Anyone remember a sci-fi movie called "The Forbin Project" (Iu think thats right). Postulates a super computer that is given control of all nukes in the US. Ends up deciding that humans are not capable of making correct decisions. Gets its way by threatening to launch nukes on selected US cities if not obeyed. Think about it!. If it had( or could force) suitable sensors to be installed in every silo,could force its humans slave to ensure it had continuous power etc, there would be no way to shut it down without causing a nuke holocaust.In the movie , it has a general executed trying to shut to it down. Ends up forcing humans to connect it to a similar Russian super computer and they take complete charge.
AI gone overboard. Can't happen you say!. How would you prevent it!

Yes, the story was from a series of three books. The movie was the first book.

 

Codifying any laws ... and the question is who makes them ... they are still implemented by the weakest link .... the human. Humans still falter to the various vices.

Its like denying all the good that humans have achieved. In a way, the medium you use to make your point also has two sides. You flip the coin and there are fake news and other issues that come with the trade. But we never say its a bad thing.
AI is going to be the same, but with greater implications thus, the need to regulate and create transparency guidelines.

 

Its like denying all the good that humans have achieved. In a way, the medium you use to make your point also has two sides. You flip the coin and there are fake news and other issues that come with the trade. But we never say its a bad thing.
AI is going to be the same, but with greater implications thus, the need to regulate and create transparency guidelines.

Absolutely, but as someone has already said, rules are made to be broken and due to the greater implications of AI is a simple set of guidelines enough? I am not suggesting this shouldn't be a start, but perhaps more needs to be done?

 

Its like denying all the good that humans have achieved.

Not denying anything. You already read all the "good", but the flip side needs to remind people that HUMANS will do for them individually .... and yes, that is more an ethical issue vice a codification of laws issue.

One could write a law, say the 14th Amendment to the Constitution, where the authors intent was well known, yet wasn't specifically expressed in the amendment, whereas now the author's intent is cast aside.

Or take the law titled "Inattentive Driving." Every jurisdiction has written a law about not texting and driving. There was no correlation between Inattentive Driving and texting while driving.

So yes, the weakest link is the human.

Humans can be outstandingly good or outstandingly evil. You don't get one without the other.

 

So yes, the weakest link is the human.
Humans can be outstandingly good or outstandingly evil. You don't get one without the other.

I am optimistic that the tech could be outstandingly supportive and assist us to take care of the inattentive driving part.
The way I see it, we can achieve miraculous options only we give it enough time and testing.
The intent can always be questioned, but if we progress towards AGI, the chances to resolve conflicts would improve more than abuse.

 

Absolutely, but as someone has already said, rules are made to be broken and due to the greater implications of AI is a simple set of guidelines enough? I am not suggesting this shouldn't be a start, but perhaps more needs to be done?

Guidelines don't come handy only during development, but also to keep a check where you deviated from the trajectory. There will be glitches and disruptions but, like your question itself answers, the 'implications of AI'/ potential is immense and we need to be optimistic to find solutions as we tread along.

 

I am optimistic that the tech could be outstandingly supportive and assist us to take care of the inattentive driving part.

Inattentive driving has always been with us. Before the radio was installed in a car, you had passenger or passengers. There probably isn't a young man who hasn't been distracted by a young woman, whether they were driving, walking, or riding.

Now, when you install the "AI driver," that only allows the "responsible" driver the illusion of safety. Given enough time ... well, given enough time people will learn to live together. Hell, we only been attempting that little feat since mankind first walked on this planet.

History is replete with examples of those who choose a different route, vice living together in peace.

I'm not anti-AI. I just have a difficult time believing it will be "good" in all situations. Now, if we have "AI" drivers, who will be the payee for accidents? Hell, I wasn't driving, Auto was ... Automatic.

 

Inattentive driving has always been with us.

Exactly, so why not give AI(that doesn't get distracted) a chance?

Now, when you install the "AI driver," that only allows the "responsible" driver the illusion of safety. Given enough time ... well, given enough time people will learn to live together. Hell, we only been attempting that little feat since mankind first walked on this planet.
History is replete with examples of those who choose a different route, vice living together in peace.

I won't say I differ on this point but you must reckon that technology has been part of our evolution. Whether we went to wars or signed peace treaties, technology has always been there. You cant exclude it.

I'm not anti-AI. I just have a difficult time believing it will be "good" in all situations. Now, if we have "AI" drivers, who will be the payee for accidents? Hell, I wasn't driving, Auto was ... Automatic.

And I don't think, AI is 'good' in all situations, either. I am against the mere thought of using AI as warfare/modernize weapons. No, it must never be applied to these fields. But AI as self driving cars seems ok to me.

A good point you raised is regarding insurance. This is my area of interest too. Nothing credible yet on this. Whatever few mishaps that will happen(if at all they happen), might be covered by the car manufacturer, but they may not agree and blame the developers. So this would be interesting to see.

 

If what we had was actually AI then laws directed at the acting agent would be needed. What we have instead today is a hammer that can be swung in 10 high dimensional space to represent the image of a scooter. The only intelligence in the system is from the humans that identified scooters in the original images used to build the classifier.

https://www.flickr.com/photos/129821058@N05/

 

Whatever few mishaps that will happen(if at all they happen), might be covered by the car manufacturer, but they may not agree and blame the developers. So this would be interesting to see.

It's not "if at all they happen" because they already have happened.

Arizona is but on case. That case alone will give politicians a pause before deciding on allowing AI driven cars.

One can make the case it's only a minutia of errors, and I agree. Once the liability situation is answered, and someone's bank account is out a few 6 or more figures, the problems might get resolved.

As far as the military goes, not much happens in society that has not already happened in the military. Again, reducing people, the highest cost line item, is typically a goal that appeals to most decision makers.

 

It's not "if at all they happen" because they already have happened.

I was referring to the more comprehensively tested and accurate vehicles that will roll out in future.
What happened in the past is a clear case of negligence. You cant bring out a product in real life situation when its not 100% tested and checked for millions combinations of scenarios.

One can make the case it's only a minutia of errors, and I agree. Once the liability situation is answered, and someone's bank account is out a few 6 or more figures, the problems might get resolved.

Maybe you are right. Once these innovators and manufacturers are brought under pressure and stood liable, I am certain they will ensure the products are accurate and deliver their best performance.

 

I see it as a loosing battle because like always the "REDACT"ing guy with the money that wins!

There is no reasoning. Just the"REDACT"tards trying to hustle your money, thought or time.