August 23, 2023 12:31 PM

Do Machines Have a Right to Read?

Yes, according to Jeff Jarvis. That is one of his unpopular opinions about AI:

Machines should have the same right to learn as humans; to say otherwise is to set a dangerous precedent for humans. If we say that a machine is not allowed to learn, to read, to extract knowledge from existing content and adapt it to other uses, then I fear it would not be a long leap to declare what we as humans are not allowed to read, see, or know some things. This puts us in the odd position of having to defend the machine's rights so as to protect our own.

I have been staying mostly out of the AI debate these days, except as devil's advocate inside my head to most every arguments I see. However, I must admit that Jarvis's assertion seems on the right trackto me. The LLMs aren't memorizing the text they process so much as breaking it down into little bits and distributing the bits across elements that enable a level of synthesis different from regurgitation. That's why they get so many facts wrong. That sounds a lot like what people do when they read and assimilate new material into their heads. (Sadly, we get a lot of facts wrong, too.)

My support for this assertion rests on something that Jarvis says at the end of his previous bullet:

I am no lawyer but I believe training machines on any content that is lawfully acquired so it can be inspired to produce new content is not a violation of copyright. Note my italics.

I, too, am not a lawyer. Good read from Jarvis, as usual.

Posted by Eugene Wallingford | Permalink | Categories: Computing