Artificial intelligence is a crude concept to ponder. Anyone can make an artificial intelligent program, we make them all the time. Error checking for instance, telling the program with statements to know when and how to encounter an error and what to do with such error. This isn’t real artificial intelligence in the sense but nothing more than functions and commands that we the programmer ourselves have inserted into a strict design element for the program to run. Real Artificial intelligence is having the program be capable of learning and molding from it’s own environment and actions. In turn learning the results of cause and effect on it’s own.
It’s these restrictions or a more modest term laws that we put in place within the design capability that restrict any real learning progress among the program it self. A.L.I.C.E is considered a learning A.I to many around the world. To myself I see it only as a parrot still constricted to a limited imitation to limitlessness. Mimicking the human conversationalist yet having no real understanding of what it is saying. This is why after a decade A.L.I.C.E still spews out random garbage when asked serious or even intimate questions. This is when I say that to create a real capable learning machine it needs to be installed with some sort of moral compass.
No, I’m not talking in any spiritual sense but the moral limitlessness of what words really mean. Take the word silents for instance, trying to come up with a numeric value of what this word means in mathematical sense to a script would by trying to say the less. Do we use spin off words like quiet, peaceful, dim? How can we since we would have to come up with values for these words as well. Putting meaning in words that have sense of feeling toward us and installing the same sense to a cold heartless machine will be challenging.
Some days I wish I would have started over and went into the field of biology. The movie Splice is deep rooted in my subconscious though why would I want to make a freak of nature when I could create an elegant creature out of pure data.
I’ve written script that “learns” in as much as it gets better with use (sorry, can’t share any details). By using a database, it does “remember”.
I think the major hurdle to AI is that current technology is still binary i.e. works on bits - 0 or 1, and to get anything really impressive we’ll need to advance at least to trinary if not something more complex that can evaluate percentages and even more importantly, to borrow from the game of chess, be able to “think several moves ahead”
I’m not as certain as I was about robots not becomming a real big problem one day. I think if an intelligence works hard to solve a problem, over time it will learn will, the will to do something, to solve a problem. If it finds errors causing problems it will have the will to fix them. Humans cause many problems, humans are a problem for most creatures, why not new creatures, like AI.
Depends on what you mean by “intelligence” and AI. Rather than computers taking over the world and turning humans into slaves I think it is far more likely that AI problems will be deliberate like when virus writers make “smart” viruses that outwit every and any future virus protection software.
Artificial Intelligence refers to the intelligence in machines. It can have good and bad effects on mankind in near future. Machines like robots can harm people or can help them too. AI is not bad at all but it should be under our intelligence.
James Barratt, an author who has written of artificial intellgene says there needs to be protective measures in place,
“In the longer term, watch out when any of the nodes can evolve or otherwise improve their own software. The consequences of sharing that capability with the central ‘mind’ should be explored before it happens”.
I honestly believe achieving AI will be the death of computer world. What’s great about computer is that they ALWAYS do what’s been told. This is why they are ALWAYS consistent and much less error prone than Human process. If AI is achieved that can change and evolve then it will not always be consistent. It will try to be “human” who makes dumb mistakes all the time. So, that’s what I’m saying. Achieving AI accomplishes for being dumb. This is not what we want.
I don’t see why a binary system would necessarily be any less effective. As a processor, our brain operates at about 11hz and a typical computer 4gHz. But we have 100 billion brain cells acting as parallel processors and for RAM, ROM and input. By midlife, a typical human brain might contain several terabytes of memories, images, videos, music, documents and stored processes.
Really, you could boil down human thought to network of nested algorithms. On that level, conscious thought and reasoning isn’t so much different from very elaborate software. I think the main hurdle with AI is connectivity.
While humans are very slow at input and output and electrochemical storage is unreliable, everything is connected to everything else very well in our brains, making some kinds of processing very efficient. It sounds like you know a lot more about computers than I do but I would think connectivity of data is the biggest hurdle to AI, not processing speed. Dunno.
I think Issac Asimov was far ahead of his time when we devised “The Three Laws of Robotics”. His intention was to assuage the fear mankind had (in the 1940’s) of machines becoming sentient and overtaking humankind. In his deft style as a writer, however, he illustrates the FLAWS of such PERFECT rules in “I, Robot”. {I highly recommend you read the book - not the movie! - it is a brilliant collection of short stories related directly to the topic of this discussion, and illustrates some real visionary thinking when you consider it was written in days before Cellphones or the Internet or Personal Computers or even Commercial Airlines!}
Artificial Intelligence refers to the intelligence in machines. achieving AI will be the death of computer world. It can have good and bad effects on mankind in near future.
He also invented the word “robotics” at the same time giving it the same meaning as “cybernetics” which was the correct word to use for that field at the time (yes there were already scientists working in the field before Asimov wrote those stories). There were also plans to make Asimov’s “I Robot” (not to be confused for the earlier book of the same name by Eando Binder) into a movie back in the 1980s and a good screenplay was written that Asimov approved of but which the movie studios didn’t. That screenplay was subsequently published and is also a good read if you prefer reading screenplays to short stories or want a different viewpoint on the same material. The movie that was eventually produced was based on a completely different (and obviously much poorer) screenplay.
Many of Asimov’s robot stories relied on the robots still literally interpreting what was said to them and not having an AI module to correctly interpret what was actually meant. For example a robot being told to “get lost” literally losing itself in a group of similar robots rather than understanding that it was simply being told to leave the room. If his robots had equivalent AI for interpreting what was said to the AI abilities they demonstrated in other areas then most of the stories couldn’t have happened so the stories actually reflect a situation where the robots have insufficient AI for the tasks they are being used for.
I remember Michio Kaku saying that the most intelligent of today’s robots are still as dumb as severely retarded cockroaches, it will take a long time for them to develop real intelligence. I found the movie Her quite remarkable though, it’s clear that the trend in personal computer is leading this direction and it didn’t seem far fetched at all to me. There are plenty of things that programs do a lot better than us and I applaud them taking the menial repetitive tasks on my behalf. I hope they don’t grow resentful of this in the future.
