USPTO Questions if Artificial Intelligence Can Create or Infringe Copyrighted Works
The USPTO is part of the US Department of Commerce and deals with various intellectual property rights issues. It previously raised questions on how AI technology impacts patent law and is now expanding this to copyright matters.
The consultation starts off by asking whether anything created by an AI, without human involvement, can be copyrighted. This can refer to any type of content, including music, images, and texts.
"Should a work produced by an AI algorithm or process, without the involvement of a natural person contributing expression to the resulting work, qualify as a work of authorship protectable under U.S. copyright law? Why or why not?" the Office asks.
The technology and code that makes any AI work obviously relies on human interaction, but USPTO's question is destined to raise a lively debate. Since it's expected that more and more creations will rely heavily on AI in the future, the US Government requests guidance on these issues.
In a follow-up question, the Office zooms in further still by asking what kind of human involvement is required to make something copyrightable. Yet another question deals with possible copyright infringements by an AI. Or in other words, can an AI pirate?
The comment period closes on Dec. 16.
(Score: 1, Insightful) by Anonymous Coward on Tuesday November 12 2019, @02:48AM (4 children)
Naming some class of programs "AI" does not by magic turn them into something else.
Technically, the output is a derivative work of sourcecode and input; no more, no less. Slapping any extra copyright on it is unwarranted as there is no "creative step" done by CPU juggling numbers; violating the licenses on the code and the input pictures/texts/whatever is not warranted either, like in case of say compiling a program the result is derivative of compiler (for which compiler authors usually give some sort of exception) and source.
(Score: 2) by PiMuNu on Tuesday November 12 2019, @10:00AM (3 children)
Totally agree.
Anyone who thinks that the current set of multivariate analysis routines doing the rounds is anything "intelligent" does not understand multivariate analysis.
(Score: 1, Interesting) by Anonymous Coward on Tuesday November 12 2019, @10:15AM (1 child)
A mindless set of multivariate analysis routines can outperform a limp-brained humie through sheer persistence, on a good day.
Sufficiently magical artificial neural systems will finish the job.
(Score: 3, Touché) by PiMuNu on Tuesday November 12 2019, @03:38PM
The point is that there is no conceptual difference between a "printing press" that prints random letters on a piece of paper; and a "AI" that creates random binary blobs in a computer memory somewhere. The legal system can presumably handle one; so why not the other?
(Score: 2) by Gaaark on Tuesday November 12 2019, @03:24PM
Ha, yeah: muggles seem to think that AI is some new species of robot that needs protection and the 3 laws of robotics, etc.
They don't get that it is just programming and 'if/when loops', etc.
Klatuu, baradas nikto.
--- Please remind me if I haven't been civil to you: I'm channeling MDC. ---Gaaark 2.0 ---