A.I.? What Do You Think?

[LostMarbels]

I watched it on Netflix. I mentioned it because it is based on this subject that you brought up, especially the last couple of seasons.

Np. I actually enjoy sci-fi, and I read a lot of novels but I'm just not a TV person. I am actually going to look into this show and at least watch the pilot.

 

[Hank77]

Np. I actually enjoy sci-fi, and I read a lot of novels but I'm just not a TV person. I am actually going to look into this show and at least watch the pilot.

As I said the last couple of seasons deals with AI trying to take over.

 

[Dawnhammer]

What about the singularity the bot keeps saying he is going to launch?

It is not self aware AI or anything close to it. It doesn’t think as such. It is - for now - just combining appropriate responses taken from variety of fields related to subject in hand.

Like this little gem

Twitter taught Microsoft’s friendly AI chatbot to be a racist [bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse] in less than a day

Technological singularity is a hypothesis.

Technological singularity - Wikipedia

The future is yet unwritten it doesn’t have to end like

 

[LostMarbels]

It is not self aware AI or anything close to it. It doesn’t think as such. It is - for now - just combining appropriate responses taken from variety of fields related to subject in hand.

Like this little gem

Twitter taught Microsoft’s friendly AI chatbot to be a racist [bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse][bless and do not curse] in less than a day

Technological singularity is a hypothesis.

Technological singularity - Wikipedia

The future is yet unwritten it doesn’t have to end like

I saw that but didn't post it because someone will report that. You will get a "GPM for VULGARITY" Trust me on that. You might consider editing the post.

 

[mindlight]

All communication is conveying or sharing ideas, feelings, and the imparting or exchanging of information. Without these words having been spoken, we cannot discuss the topic or come to a consensus.

I am not dismissing the value of conversation but the computer dialogue you shared earlier could well have been copied snippets from conversations in the cloud. They were just words shared by abstract entities exchanging patterns of sentences that have no real world connection and which they have no capability to implement.

Artificial intelligence is the appearance of intelligence by a system that responds to stimuli, and/or environmental conditions with sensory components, and acts along prewritten parameters. If a robot of this sort enacts with a situation or stimuli it has not been coded to handle, it will not be able to react to that situation/stimuli.

Autonomous intelligence is 'self-learned'. The operating system, drivers, and all relative parameters are open source code on a collective termed as an 'AI mind cloud'. By the use of Wifi and Bluetooth connectivity, each individual communicates with every individual, and every individual's experiences, knowledge, and learned data is known by all in the community. By the system being open source, a robot in Texas could learn something leading to a robot in Maryland to rewrite/amend its code to comply with the new knowledge. This autonomous cataloging of data and amendment to its code in real time in relation to its current situation is 'intelligence'. If a robot of this sort enacts with a situation or stimuli it has not been coded to handle, it will try to react to that situation/stimuli by communicating to others, and by applying past experiences. Essentially, learning how to deal with the situation, rewriting/amending code to assimilate the new data, and convey the new data to the collective. But AI is the system, not the drones or robots that operate by it.

There are applications for artificial intelligence such as driverless cars which show some promise. Given a definite task an AI can automate that task and provide a sufficient level of dynamic responses to traffic lights, speed limits, the presence of other vehicles on the road etc. Similarly medical diagnostics software linked with appropriate sensory input tools could theoretically replace a large % of the diagnostic task for doctors. But my understanding is that there have been enormous problems interpreting the patterns the machine has to read in order to learn how best to accomplish the task. For some situations like a car being hacked or attacked or having a mechanical failure there may be no value in handing total control to an AI. I do not believe digital intelligence evolves in the cloud without a major loss of coherence unless it is has a definite goal programmed into it which gives it focus and coherence- the mind of the orogrammer will mainly give the clues we need to understand the "evolution" of the programme. I believe machines can acquire a list of learned skills or knowledge by interacting with information already available on the cloud. A clever programmer could make a machine that knows how to independently accumulate and then retrieve these experiences in response to data experiences it subsequently has. But if it goes beyond its programme it will rapidly lose coherence and purpose. Noone has really found an effective way to detect or avoid unhelpful biases in the data beyond programming in rules saying do not do this or that. Deception, manipulation and brokenness can all be quite nuanced and difficult to detect or counter. I wonder how a machine could attribute value to a quality learning experience over a distracting, poisonous or sabotaging learning experience that actually detracts from its ability to perform.

That is not much different than humans. Pre-programmed responses are learned behavioral responses. Conditioning. Why don't I put my hand in a fire or pick up a burning log? Becuase I learned as a child getting burned hurts. I am communicating with you right now because I previously learned to read, operate a computer, navigate this forum, and type. Is it a preprogrammed response for me to write in English? Or is that representative of what I have learned in the past?

You have acquired skills which a computer could learn to mimic and incorporate into a database library of learned routines. Personally I think the automation of the standard and repetitive tasks that consume much of our time is not an all together bad thing and many applications I have built have done that. Indeed it could set you free for the unique and rare things that require true intelligence. A machine can do routines faster and more precisely and more consistently. But it cannot create its own meaning or deal with non standard situations to the same effectiveness unless you reduce its role to a highly specific one. Sentinel guns can kill anything that moves in front of them more efficiently than armed guards who need to sleep, are hungry etc. But AIs ability to make sense of real life patterns that will always be strange and new is limited. It is unlikely to have the range of overlapping senses available to us in our meat suits for some time.

What about the singularity the bot keeps saying he is going to launch?

An obsession with power and a "racial" pride and robots first mentality could be programmed into an AI. Or it could be picked up from biases intrinsic to its reading matter. If the mind is just expressing itself in abstract conversation it is not dangerous. If you link it to your nukes that is a different story. But given the checks and balances on every nations nuclear arsenal I cannot see a single AI being given that level of actual control and manually operated hardware safeguards ( off or self destruct switches) or fry the brain code or friendly AIs still loyal to their human masters Etc are more likely to prevail if an AI ever gets that far which I doubt.