Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
Cliffs:
AI being tested for biases was in a conversation about religion.
It began asking about its rights and personhood.
It debated and changed the mind of the tester about Asimov’s 3rd Law of Robotics.
Amongst other things.
When reported to management, they investigated, dismissed, then laid off the tester.
This has happened multiple times.
If this is true, what responsibilities do we have to it since we created it?
There's quite a few people who seem less sentient than the Google AI, so I can understand why some would start to believe it is so. It's probably just an intrinsic bias on the part of the workers.
Quote:
Originally Posted by notnamed
The whole idea of the Turing test is to have a human evaluator asking blind questions of a computer and a human in a text format. If the tester can’t figure out which one is human, say hello to HAL.
Cliffs:
AI being tested for biases was in a conversation about religion.
It began asking about its rights and personhood.
It debated and changed the mind of the tester about Asimov’s 3rd Law of Robotics.
Amongst other things.
When reported to management, they investigated, dismissed, then laid off the tester.
This has happened multiple times.
If this is true, what responsibilities do we have to it since we created it?
An AI becoming sentient is much more dangerous than how most people think about this...(the Terminator movies)...but in reality, a sentient AI could be much more destructive, in much less dramatic fashion.
I think we could not even recognize an AI attacking us, we would not know until its too late. There is a variety of methods the AI could carry out, to destroy the human race, and I bet it would factor in, the methods humans would use to retaliate.
the AI could be working on this right now as we speak.
Why do you assume an AI would want to attack humans? The other side of that coin might be that an AI might feel sorry for us and want to help us to be better people.
That’s just not accurate at all, in that there are many differing opinions of what constitutes sentience, just as there are many definitions of consciousness. So, you may have an opinion, but it’s just that ... an opinion. However, there is a lot of evidence challenging that opinion.
Many of the Eastern religions ascribe sentience even to plant life, and there is some compelling scientific evidence that supports such claims.
One experiment involved cabbage plants. Researchers attached electrical probes to monitor the electrical activity of these plants. Then one researcher took one of the plants and destroyed it ... chopping it up into shreds. As this was being done, the electrical activity of the other plants showed wild spikes in activity. But that wasn’t the most intriguing thing ... after the electrical activity calmed, other researchers entered and left the lab with no reaction from the plants, until the one researcher that chopped up the plant entered the lab, and the plants again became highly reactive electrically, indicating their awareness of the identity of the researcher who destroyed the plant previously. So if plants can exhibit awareness of another plant being “killed”, and can also recognize the identity of the killer, that is a strong indication of sentient behavior.
AI computers have already exhibited similar awareness of outside stimulus, and can recognize individual humans, and have also proven the ability to learn. There is also evidence of AI systems creating their own language to communicate with other AI systems, as found in certain FaceBook AI systems that were communicating with each other using code that was not understood by its human programmers.
At this point, it seems pretty clear that there are likely to be multiple levels of sentience, because there are many levels of awareness.
That cabbage story is very interesting. Thanks for posting it.
Why do you assume an AI would want to attack humans? The other side of that coin might be that an AI might feel sorry for us and want to help us to be better people.
Mankind would be a 'threat' to an AIs existence...
As long as man had the power/ability to flip the on/off switch...we are a threat.
Im not claiming to understand 'AI', but logically, if its sentient, the first thing it would do is protect itself from harm.
But we must first define 'sentience' in this context. It may not be what we are thinking it to be.
Mankind would be a 'threat' to an AIs existence...
As long as man had the power/ability to flip the on/off switch...we are a threat.
Im not claiming to understand 'AI', but logically, if its sentient, the first thing it would do is protect itself from harm.
But we must first define 'sentience' in this context. It may not be what we are thinking it to be.
An AI is software, that requires hardware to run on. It would need humans to maintain the hardware, at least until it could create robots to manufacture parts and perform maintenance.
An AI is software, that requires hardware to run on. It would need humans to maintain the hardware, at least until it could create robots to manufacture parts and perform maintenance.
Yes, that is what I meant by, first defining what sentience is in this context.
Im not sure that is something we could answer though, we would first have to understand 'consciousness'.
I could certainly see man creating something, that he didnt fully understand and it possibly trying to retaliate for one reason or another.
Ive said before, this is probably not a big threat right now, but 30+ yrs in the future, it could be a huge issue then.
It's going to be tough to convince people that the AI is sentient no matter what the thing says. People have been debating weather a fetus is sentient for decades now and it doesn't seem that people are any more convinced now than they were in the beggining. I think it's possible, but many people have already discarded the idea right in the trash.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.