Page 2 of 3 FirstFirst 123 LastLast
Results 31 to 60 of 85

Thread: AI patents

  1. #31
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by profloater View Post
    Nice reference, it illustrates the problem. There is an algorithm, the human input with the human objective. But no human predicts the result. The result is useful, can it be patented? I would expect the algorithm to be challenged as using prior art, or even obvious to one skilled in algorithms. But without patent protection, will the investment be made? Or is it open season for labs with robots?
    If the bot is property of a company, the owner will claim the patent. I predict no serious legal challenges.

    What do you mean by "open season"?
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  2. #32
    Join Date
    Jun 2005
    Posts
    14,430
    Quote Originally Posted by Noclevername View Post
    If the bot is property of a company, the owner will claim the patent. I predict no serious legal challenges.
    I would agree with that. In fact, typically when scientists use some AI program or something similar to do experiments, there is a prior agreement (memorandum of understanding) signed between the institutions on how the intellectual property derived from the experiments will be shared.
    As above, so below

  3. #33
    Join Date
    Feb 2005
    Posts
    11,950
    I recently read a paper by Andrew Cooper called “A Mobile Robotic Chemist” that may point the way forward in many fields.
    ISS should have a few.

  4. #34
    Join Date
    Feb 2003
    Location
    Depew, NY
    Posts
    12,217
    Quote Originally Posted by Noclevername View Post
    If the bot is property of a company, the owner will claim the patent. I predict no serious legal challenges.

    What do you mean by "open season"?
    I suspect someone will get the idea of creating an AI to patent AI discoveries faster than the human ever could. I could see that character getting into a lawsuit for their troubles. An AI to navigate a website has to be easier than an AI to navigate Mars. You know there is a schemer, with really bad ideas out there.
    Solfe

  5. #35
    Join Date
    Mar 2010
    Location
    United Kingdom
    Posts
    7,311
    Quote Originally Posted by Solfe View Post
    I suspect someone will get the idea of creating an AI to patent AI discoveries faster than the human ever could. I could see that character getting into a lawsuit for their troubles. An AI to navigate a website has to be easier than an AI to navigate Mars. You know there is a schemer, with really bad ideas out there.
    Hey, just because I suggested it in post 6 doesn't mean I am an evil schemer.

  6. #36
    Join Date
    Apr 2011
    Location
    Norfolk UK and some of me is in Northern France
    Posts
    9,228
    Quote Originally Posted by Noclevername View Post
    If the bot is property of a company, the owner will claim the patent. I predict no serious legal challenges.

    What do you mean by "open season"?
    So long as so called AI does not author the patent, I am sure you are right. Although it may be that the algorithm is invented by a third party. I could still see objections if the algorithm is shared or obvious. The invention becomes a discovery. With a discovery, first to publish is still important but there is no inventive step. I still see problems ahead or should I say opportunities for patent lawyers? With larger companies able to use larger AI or data handling systems, it increases the advantage of large companies over small. However the huge online sharing favours small users of published patents countering that advantage. And that ironically can use searching bots. So advanced data handling, a better term than AI I think, cuts into patents both ways.
    sicut vis videre esto
    When we realize that patterns don't exist in the universe, they are a template that we hold to the universe to make sense of it, it all makes a lot more sense.
    Originally Posted by Ken G

  7. #37
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    I use the term AI in its current, information science meaning. AI may not be conscious but we're learning how it can be "creative"... in a totally non-human way, of course. See https://en.wikipedia.org/wiki/Synthetic_media
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  8. #38
    Join Date
    Apr 2011
    Location
    Norfolk UK and some of me is in Northern France
    Posts
    9,228
    Quote Originally Posted by Noclevername View Post
    I use the term AI in its current, information science meaning. AI may not be conscious but we're learning how it can be "creative"... in a totally non-human way, of course. See https://en.wikipedia.org/wiki/Synthetic_media
    The reference is about fakery such as using Illustrator to falsify images. It is not close to invention as understood by patent examiners. However, if you teach a current system how to write a patent, for example by making it read good patents, there is no doubt it could generate new patents using keywords or even half baked human ideas. Circuit diagrams can be studied by AI and corrected or optimised by automatically simulating the response of the circuit, to give one example. Could an examiner necessarily detect an automated patent application? The point of this thread is to applaud the US patent office In principle but to wonder whether it can hold the line about ownership of the single inventive step.
    sicut vis videre esto
    When we realize that patterns don't exist in the universe, they are a template that we hold to the universe to make sense of it, it all makes a lot more sense.
    Originally Posted by Ken G

  9. #39
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by profloater View Post
    The reference is about fakery such as using Illustrator to falsify images. It is not close to invention as understood by patent examiners.
    No, it is just one example of machines manipulating information to produce new results without direct human input. I don't know anything about patent laws, so I can't comment on that part.
    Last edited by Noclevername; 2020-Jul-19 at 08:50 PM.
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  10. #40
    Join Date
    Jan 2002
    Location
    The Valley of the Sun
    Posts
    9,843
    Maybe patenting will become obsolete when our AI invents all the best new stuff and just gives it to us.

  11. #41
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by Chuck View Post
    Maybe patenting will become obsolete when our AI invents all the best new stuff and just gives it to us.
    That would depend more on massive changes in laws and economy than on developments in technology.
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  12. #42
    Join Date
    Apr 2011
    Location
    Norfolk UK and some of me is in Northern France
    Posts
    9,228
    Maybe knowledge worth protecting will go secret again. It is possible and is perhaps a trend with encryption progress
    sicut vis videre esto
    When we realize that patterns don't exist in the universe, they are a template that we hold to the universe to make sense of it, it all makes a lot more sense.
    Originally Posted by Ken G

  13. #43
    Join Date
    Jan 2002
    Location
    The Valley of the Sun
    Posts
    9,843
    Quote Originally Posted by Noclevername View Post
    That would depend more on massive changes in laws and economy than on developments in technology.
    It won't be necessary to repeal patent laws. People will lose interest in patents when no one wants to buy human inventions.

  14. #44
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by Chuck View Post
    It won't be necessary to repeal patent laws. People will lose interest in patents when no one wants to buy human inventions.
    I'd file that under "possible but not plausible". Some human will always want to claim the money.
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  15. #45
    Join Date
    Jan 2002
    Location
    The Valley of the Sun
    Posts
    9,843
    There would be no money to claim. No one would want their inventions when the AI has already produced something better and doesn't care about patents.

  16. #46
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by Chuck View Post
    There would be no money to claim. No one would want their inventions when the AI has already produced something better and doesn't care about patents.
    The AI's inventions (and the resulting income) will be claimed by someone. Bet on it.
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  17. #47
    Join Date
    Apr 2011
    Location
    Norfolk UK and some of me is in Northern France
    Posts
    9,228
    The whole point of the patent deal between society and the inventor is to protect investment by granting limited exclusivity to recoup the investment and make profit. It also gives the government a chance to grab ideas that are considered strategic. The motivation for any true AI has yet to be formulated outside of SF, but motivation , I think that could be called a feeling. If AI is given or evolves feelings, then the conflict between secrecy and sharing will pass from humans to AI.

    So if the patent office just banned algorithms as inventors, there would be no argument. There is no AI yet as has been pointed out, but the debate and effort to achieve AI is not anywhere near a conclusion.
    sicut vis videre esto
    When we realize that patterns don't exist in the universe, they are a template that we hold to the universe to make sense of it, it all makes a lot more sense.
    Originally Posted by Ken G

  18. #48
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by profloater View Post
    The whole point of the patent deal between society and the inventor is to protect investment by granting limited exclusivity to recoup the investment and make profit. It also gives the government a chance to grab ideas that are considered strategic. The motivation for any true AI has yet to be formulated outside of SF, but motivation , I think that could be called a feeling. If AI is given or evolves feelings, then the conflict between secrecy and sharing will pass from humans to AI.

    So if the patent office just banned algorithms as inventors, there would be no argument. There is no AI yet as has been pointed out, but the debate and effort to achieve AI is not anywhere near a conclusion.
    By "true" AI, you mean conscious AI? I meant AI as currently used in computer science.
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  19. #49
    Join Date
    Apr 2011
    Location
    Norfolk UK and some of me is in Northern France
    Posts
    9,228
    Quote Originally Posted by Noclevername View Post
    By "true" AI, you mean conscious AI? I meant AI as currently used in computer science.
    Years ago I remember writing a proposal for an “intelligent thermostat” that learned a pattern of use. I may have referred to Artificial Intelligence but never meant it to be conscious. Now we have learning systems that can cycle through millions of experiments to find new catalysts, as mentioned earlier. So I guess AI has many meanings within computer science. Similarly creativity or invention can have many uses as words but must be tightly defined in patents. When I noticed the news about the patent office, the new rule directly implies that AI is not inventive and is not expected to be, by that office, even if it has a go at invention!

    I guess the examiners will have to devise a Turing test to see if an applicant is Human, or Human enough. It would be irresponsible although tempting to have a go at designing a computer based Text and image generator to try to get through the process.
    sicut vis videre esto
    When we realize that patterns don't exist in the universe, they are a template that we hold to the universe to make sense of it, it all makes a lot more sense.
    Originally Posted by Ken G

  20. #50
    Join Date
    Jun 2005
    Posts
    14,430
    Quote Originally Posted by Noclevername View Post
    By "true" AI, you mean conscious AI? I meant AI as currently used in computer science.
    I think that is the interesting issue. At present, AI doesn’t ask to be given rights. But what happens if an AI becomes sentient and starts demanding a share?


    Sent from my iPhone using Tapatalk
    As above, so below

  21. #51
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by profloater View Post

    I guess the examiners will have to devise a Turing test to see if an applicant is Human, or Human enough. It would be irresponsible although tempting to have a go at designing a computer based Text and image generator to try to get through the process.
    If a person could plausibly design a method of fooling a Turing test using existing processes, doesn't that render the entire Turing test useless?
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  22. #52
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by Jens View Post
    I think that is the interesting issue. At present, AI doesn’t ask to be given rights. But what happens if an AI becomes sentient and starts demanding a share?
    A conscious non-human entity demanding its own rights or ownership would be very interesting to see the results, it might impact other areas such as Corporation "personhood".
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  23. #53
    Join Date
    Jun 2005
    Posts
    14,430
    Quote Originally Posted by Noclevername View Post
    A conscious non-human entity demanding its own rights or ownership would be very interesting to see the results, it might impact other areas such as Corporation "personhood".
    Yes, for sure. It would be a very interesting problem for us to confront.
    As above, so below

  24. #54
    Join Date
    Apr 2011
    Location
    Norfolk UK and some of me is in Northern France
    Posts
    9,228
    Quote Originally Posted by Noclevername View Post
    If a person could plausibly design a method of fooling a Turing test using existing processes, doesn't that render the entire Turing test useless?
    In relation to patents the Turing style test is not useless but difficult. An examiner only has the words and pictures. Most human versions would include personal interaction. We already know a typical person cannot detect an actor behind the scenes, being overwhelmed by the appearance of the person delivering the responses. The examiner has no chance. So in relation to the first post I fear the ruling is well intended but very hard to apply. So the office will refuse patents claimed by a self identified AI but surely that is not the worry? Well, I suppose newly conscious AI will worry about that! Like a patent examiner, will any of us know when that threshold is passed? We are having enough trouble evaluating bots!
    sicut vis videre esto
    When we realize that patterns don't exist in the universe, they are a template that we hold to the universe to make sense of it, it all makes a lot more sense.
    Originally Posted by Ken G

  25. #55
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by profloater View Post
    In relation to patents the Turing style test is not useless but difficult. An examiner only has the words and pictures. Most human versions would include personal interaction. We already know a typical person cannot detect an actor behind the scenes, being overwhelmed by the appearance of the person delivering the responses. The examiner has no chance. So in relation to the first post I fear the ruling is well intended but very hard to apply. So the office will refuse patents claimed by a self identified AI but surely that is not the worry? Well, I suppose newly conscious AI will worry about that! Like a patent examiner, will any of us know when that threshold is passed? We are having enough trouble evaluating bots!
    As far as General AI goes, what if we get an honest one? They might refuse the Turing Test entirely on the grounds that a lie is a corruption of information. A being made of information could very well see that as perverse.
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  26. #56
    Join Date
    Jun 2005
    Posts
    14,430
    Quote Originally Posted by Noclevername View Post
    As far as General AI goes, what if we get an honest one? They might refuse the Turing Test entirely on the grounds that a lie is a corruption of information. A being made of information could very well see that as perverse.
    Can you refuse the Turing Test? I think it's probably not possible, because the refusal itself would constitute an answer, and actually one that might convince the human that the AI is actually self-aware, which would mean that the AI would have deliberately or unwittingly passed the test... In fact, simply not answering questions could be seen as an answer, and giving nonsensical ones could be interpreted as the computer deliberately trying to demonstrate that it is not self-aware...
    As above, so below

  27. #57
    Join Date
    Jan 2010
    Location
    Wisconsin USA
    Posts
    3,288
    I don't think sentient computers would survive long. I am not aware of any species that communicates as a whole. If it was a viable way for a species to survive, it would seem that there would be many species of animals that would act as one whole, the whole time. It doesn't happen that I am aware of. Computers have the ability for all computers to be interconnected. I would think it would have to be a system that thought completely independent from other systems.
    The moment an instant lasted forever, we were destined for the leading edge of eternity.

  28. #58
    Join Date
    Apr 2007
    Location
    Nowhere (middle)
    Posts
    38,003
    Quote Originally Posted by Copernicus View Post
    I don't think sentient computers would survive long. I am not aware of any species that communicates as a whole. If it was a viable way for a species to survive, it would seem that there would be many species of animals that would act as one whole, the whole time. It doesn't happen that I am aware of. Computers have the ability for all computers to be interconnected. I would think it would have to be a system that thought completely independent from other systems.
    What does "communicate as a whole" mean?
    "I'm planning to live forever. So far, that's working perfectly." Steven Wright

  29. #59
    Join Date
    Jan 2010
    Location
    Wisconsin USA
    Posts
    3,288
    Quote Originally Posted by Noclevername View Post
    What does "communicate as a whole" mean?
    All the computers in the world make decisions and communicating as if they are one entity.
    The moment an instant lasted forever, we were destined for the leading edge of eternity.

  30. #60
    Join Date
    Jun 2005
    Posts
    14,430
    Quote Originally Posted by Copernicus View Post
    All the computers in the world make decisions and communicating as if they are one entity.
    I see what you mean now. But I kind of disagree. I don't think that they do that. I mean, you could have a supercomputer doing calculations without being connected to the Internet. And also, I'm not sure why that means they would be short-lived.
    As above, so below

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •