Is that how computers work?Quote:
Originally Posted by Copernicus
ADDED: More to the point, is that how conscious entities work?
Printable View
Is that how computers work?Quote:
Originally Posted by Copernicus
ADDED: More to the point, is that how conscious entities work?
I'm not sure why you have to ask that as a question. I think it's fairly obvious:Quote:
Originally Posted by Noclevername
That is not how computers work. And it is certainly not how conscious entities work.
I wasn't sure, but I strongly suspected. It seemed more like the closing scene in the film Lawnmower Man than anything based in actual computer architecture.
I am not worried about AI working to destroy all of us. I would think that if AI started to happen it would struggle with getting along with other AI and more likely it would destroy each other AI because it wouldn't have the culture to know how to cooperate.Quote:
Originally Posted by Jens
OTOH, it would not have the competitive instincts of human tribalism to drive conflicts.Quote:
Originally Posted by Copernicus
Human tribalism instincts might be something we have to patent, if an algorithm can be written to do this for AI. So many instincts evolved into us to maximize survival.Quote:
Originally Posted by Noclevername
I guess it all depends on how their makers socialize them before releasing them into the "wild". Certainly we do not want them learning from the Internet before they have developed strong personalities! We've seen where that leads.
That is hilarious. Some of the internet trolls I have run into, must be bots.Quote:
Originally Posted by Noclevername
Statistically, yes. There are a lot of troll bots.Quote:
Originally Posted by Copernicus
https://en.wikipedia.org/wiki/Intern...Malicious_bots
Quote:
Bots may be used on internet forums to automatically post inflammatory or nonsensical posts to disrupt the forum and anger users.
Didn't know that. Take me to your leader! Haha!Quote:
Originally Posted by Noclevername
Going back to this, the bandwidth limitations alone make it seem unlikely that multiple AI would form a "hive mind".Quote:
Originally Posted by Copernicus
What if they do form a cooperation? As learning machines, they could by definition adapt to each other's presence.Quote:
Originally Posted by Copernicus
Or, There Can Be Only One: an AI that could perceive others of its kind as a threat might stymie any human attempts to build others.
Or or, there might be a surviving winner of the AI conflicts; an entity that has just taught itself by experience to be ruthless and wipe out potential threats... :doh:
How do you program in all the instincts that species have for cooperation and all the checks and balances that would be needed. I would think it would be just as hard to do this as it would be to actually make a new species of animal from scratch. Much less an intelligent species.Quote:
Originally Posted by Noclevername
Yes, it probably would be just as immensely difficult as creating a conscious being at all.Quote:
Originally Posted by Copernicus
Don't think of AI as a "species". That's a biological term. They might be digital copies of each other or they might each be a one-off, built according to totally different design architectures.
ADDED: This has become a conversation about Conscious AI itself, rather than the patent aspect. I'll drop this line of discussion and try to stick to the OP topic.
If corporations can someday own patents, perhaps a conscious being may simply establish a corporate front and register its patents through that.
A corporation can own a patent but not be the inventor. The deal is to protect the inventor but that also protects the owner. That’s why the deal is important in society. The almost opposite is publication without a patent, (that prevents a patent deal) and maybe AI would be altruistic in that way. Being barred from invention, AI just publishes! That could make human invention hard to protect, so investment in new ideas becomes unprofitable. A new angle on the question of innovation.
I misunderstood the OP as saying corporations could not legally hold patents. My bad.
As I see it, if a corporate entity can own a patent, nothing prevents an AGI from doing so by incorporating.
I could get rich by inventing artificial intelligences and selling their own patents to them.
Well now, companies are treated like individuals so could an AI be recognised as an owner? It is a different kettle from inventing. Right now you would not get far trying to incorporate with “me and my robot“ as owners!Quote:
Originally Posted by Noclevername
An AGI might have a tough battle getting recognized as a legal person in itself. Depends on a lot of legal and political factors not suited to this Forum to evaluate. What country, what leaders, what judges, etc.Quote:
Originally Posted by profloater
Agi ?
Artificial General Intelligence, a conscious machine.Quote:
Originally Posted by profloater
Thanks that’s new to me. Watch out for ASAI, artificial self aware intelligence. The next step. Not that an AI would have to be self aware to invent something, but it would have to model alternative futures in order to imagine a change. I suppose tool use is a test of that. Leading to inventing a new tool. Humans took thousands of years to add handles to flint hand axes, tens of thousands actually, And no patent! I can imagine a financial program inventing a new derivative by trial and error plus feedback. Not a nice thought unfortunately. But at least it would not be able to patent it!Quote:
Originally Posted by Noclevername
I don't know anything about the legal limits of establishing corporations under any nation's and/or international laws, but given how easily and routinely businesses create shell companies and the like, I'd bet you don't need to prove you're human to do so.Quote:
Originally Posted by profloater
Yes companies or corporations can own other companies and hide their owners in some places, (not in major countries) but the top of the pyramid has to be human owners. It’s all paperwork so a robot could get away with it just as criminals can steal identities. I am not sure that ownership would interest AI although control might, and the main issue here is recognition and protection of inventive steps, useful ones. (A patent application has to show that it is useful,) but inventors sometimes invent problems to solve.