Morality of Artificial Intelligence

  • Thread starter Thread starter Lifeisbeautiful3
  • Start date Start date
Status
Not open for further replies.
Bon_Croix said:
Robots can communicate and the communication can be made audible.
Click to expand...
Why would it be? That would be totally unnecessary and the most inefficient way they could communicate. I think some robots are trolling you, friend…
 
Last edited:
Lifeisbeautiful3 said:
AI morality is a fascinating (scary?) topic. Take self-driving cars as an example. If there’s an imminent accident, should the car veer to one side and hit a pedestrian or to the other and hit a school bus full or children. Should the utilitarian ‘good of the whole over good of the few’ trump?

What if one decision helps more people, but is likely to hurt the passengers in the car? A human driver might prioritize his/her own family over the lives of unknown strangers (even if the number of people harmed is potentially greater).

Moral underpinning Of the programmers may well influence outcomes. What if one could choose which AI morality they would like to prioritize. Who would prioritize unknown strangers over their own family
Click to expand...
We’re a long way from self-driving vehicles being able to identify collision hazards with anything like that kind of precision and accuracy. (How is it going to know who is in a bus?)

When and if we have that, I’d wager that the odds that the machine would be able to make some kind of decision would be better than the human driver who simply couldn’t think and react that fast. In the meantime, we’ve lived side-by-side with trains for a long time. What we have done is to keep people out of the path of the trains, because they can’t stop on a dime. I would think that safety decision-making will reflect the abilities of the technology and the extent to which it can be programmed, but I don’t think either/or scenarios with pedestrians and mass transit are going to be left to chance, when that can be avoided.

In other words, the best moral answer is to prevent challenging the drivers of passenger vehicles with that kind of choice as much as that is possible. Whether the driver is a fallible human or a computer whose algorithm cannot be unlimited in its subtlety, it isn’t fair to expect that kind of decision to be made on the spot. Better traffic engineering is a far better solution.
 
Last edited:
Thinker_Doer said:
If you would read up on the test results, you would see that the AI-s are much more reliable than the human drivers. When all cars will be driven by AI-s and they will all access the same central cloud database, where every car reports its position, speed and destination, along with the weather data, the road conditions, there will be no accidents.
Click to expand...
There will always be human drivers that the AI will have to deal with. And the no-win scenario as I mentioned earlier will be a possibility.
As to all cars communicating with each other…not likely. My Dodge truck will be on the road for a long time, and I can assure you it does not talk to anyone.
Information is always going to be limited, and unforseen circumstances will arise.
In those cases, AI is simply not qualified to make such a decision.
 
The latest “driverless” technology relies on object recognition via sensors, so yes, communication between vehicles, or vehicle identification via microwaves, could improve the survival odds of all nearby. Similar to the Identify Friend Foe system in aircraft. The onbard computer will make different allowances for a nearby truck than for a micro-mini car, which would not likely survive impact with a curb-side trash container without significant damage.
 
Driverless car need faster network for faster decision making.

If only they have dedicated circuit for driverless cars only, safety coillision-avoidance scenarios will be more predictable for AI.

Human drivers drive manual cars (no AI), and yet, human drivers has their sophistication in decision making faster than our 4G network.
 
Thinker_Doer said:
Technical errors can occur, but I will trust my life on AI-s rather than humans.
Click to expand...
Remember the recent Boeing 737 Max accidents? Computer isn’t always better. Specific training to operate it was needed, because computers are not as sophisticated as human in terms of decision making. Well, not yet. They are reliable for specific scenarios/ specific job scope only

They are programmable and useful. But humans decides how and what to do with them for now & future.
 
Last edited:
Current estimates are that computers will match the processing ability of humans in about 10 years. At some point thereafter, the computer will be able to compute faster than a human.

It’s only a matter of time. We will need to learn to work with computers, not independently of them.
 
Computers are just tools for people to use. They will solve problems faster but they will never be people or have the concerns of people.
 
francisca.chapter3 said:
Remember the recent Boeing 737 Max accidents?
Click to expand...
Yes, I do. They were due to the negligence of the operators, who did not implement all the necessary safeguards. So far millions of miles have been driven by the AI-s (with human supervision as it should be during testing) and there was only one accident due to a human driver’s error. How many accidents happened during the same time due to human errors?
 
The AI is good so long humans privacy is respected. This is in my opinion the focus of AI ethics.

I think it is unethical to let robots create fake accounts spreading fake news. Now in youtube there are many “compilation videos” example cute cats, cooking tips are compiled by robots. The robots get more views than the humans who make the original videos. I think these are unethical use of robots. They propagate messages as if they are real humans but not really. They are allowed to take advantage claiming other’s videos because individu personal privacy not protected enough. Google/ youtube could have known that these robots comb clean other people’s cute pet videos by making compilation of cute pet videos not belong to them, but google & youtube do nothing, because it attract more ads.
 
Last edited:
Maybe you’re a bot! Haha. Just kidding…
 
Last edited:
There were two accident related to the MCAS of Boeing 737 Max. Both because of lack of training. Both were good experienced professional pilots.
 
Thinker_Doer said:
So far millions of miles have been driven by the AI-s (with human supervision as it should be during testing) and there was only one accident due to a human driver’s error. How many accidents happened during the same time due to human errors?
Click to expand...
Red herring.
The issue is not the rate at which an AI is involved in an accident.
The issue is the moral decision that will eventually be asked of an AI involved in a traffic fatality.
 
I don’t think this doctor understands AI. What he’s referencing are medical algorithms which doctors use. Overly depending on them could be a problem but not their existence! Doctors today would be pretty lost without them.
 
I simply noted it as an aside. Should have labeled it as that Yet, the day may arrie hen AI determines the treatment. Cancer therapy, to be specific, is heavily depenent upon research data and population demographics. While treatmnt is becoming individualized, it also risks becoming “cookie cutter.”
 
Last edited:
Status
Not open for further replies.
Back
Top