[AndyTheBear]

...lower court’s ruling that patents can be issued only to human inventors and that his AI system could not be considered the legal creator of two inventions that he has said it generated.

This guys lawsuit seems ridiculous.

If somebody uses a tool, be it a piece of paper, a calculator, a pencil, or software to create something then we have always considered them to be the creator of an otherwise patentable product right?

Nobody stops to think it might be the piece of paper, the calculator, the pencil, or software that ought to be considered the creator. Why should AI software be different?

[Rusty0604]

The government figured out they would be losing out on payroll taxes with robots. I remember the proposal that companies using robots would have to pay taxes on thenm to compensate the government for their supposed loss.
I don’t know if that went anywhere. I’m sure they extract the money somewhere.

[Rebel_Ace]

"...If somebody uses a tool, be it a piece of paper, a calculator, a pencil, or software to create something then we have always considered them to be the creator of an otherwise patentable product right?

Nobody stops to think it might be the piece of paper, the calculator, the pencil, or software that ought to be considered the creator. Why should AI software be different?..."

Here's the fundamental difference, which I will illustrate with some examples...

In all these examples, some one named Joe has identified some need, but does not have a clue about how to fill it.

Example 1: Joe walks up to Frank the Inventor and asks him, "I would like a device that runs on sunlight and produces hydrogen for fuel later. Can you make that?" Frank works hard and creates the plans for the device. Who did the inventing? Obviously Frank.

Example 2: Joe walks up to Frank's office, but Frank is away at lunch. Joe asks Frank's calculator, then Frank's copy of Excel, and all the paper and pencils on Frank's desk for his 'Sunlight to Hydrogen' device. Nothing happens, because none of these things have any AGENCY to act on their own. Who did the inventing? Nobody, because no invention took form.

Example 3: Joe is tired of waiting for Frank, so he calls up a version of OMNI-Vent 3.5 AI, and asks it for his 'Sunlight to Hydrogen' device. The AI firsts looks to its training data, and finding nothing directly applicable, scours the vast internet for scraps of info that could be useful. It forms models, runs simulations, does 'competitive iterations' hundreds of thousands of times. In a future where this AI has access to physical manipulators and 3D printers, it might actual conduct real world evaluations. It then presents plans to Joe for his requested device. Who did the inventing? Clearly not Joe, he is just wishing for the thing. The programmers of OMNI-Vent 3.5? They invented an INVENTOR, so you might legally say they own the product of OMNI-Vent's output, but none of that team might know the slightest bit about chemistry or photovoltaic reactions.

The biggest misconception I see here on FR about AI, is that most people think it behaves like the simple, "linear instruction" programs they saw in college, or current use or write at work. Those are quite deterministic, where the author of the program has written specific rules for every condition the program will encounter. AI work today involves the synthesis of millions and millions and millions of test case inputs and results, where the system itself identifies patterns and conditions it deems relevant, and in many cases is free to alter its own base code. To even the people that create such systems, their behavior is not deterministic in any practical sense of the term. It in some sense has AGENCY to act.