Industry Committee on May 11th, 2018

[Kevin Williams]

If I publish a book, and I have a room of 50 people who want that book, generally speaking, 40 people will want the print book and 10 people will want the e-book.

I consider my job to be the intermediary between the author and the reader. Obviously, in addition to all the editorial production, marketing, accounting, and shipping work we do, we also try to make sure that people can access the material in the formats they want. We try to publish the books simultaneously in both print and e-book forms.

The e-book market has basically stopped growing and seems to have reached its natural limits. My own perception of it, having published books for the last 10 years and having seen how people consume them, is that, generally, you'll see this kind of breakdown: about 40 people will want the print and about 10 will want the e-book.

[Majid Jowhari, Liberal, Richmond Hill, ON]

Mr. Chair, how much time do I have?

[The Chair, Liberal, Dan Ruimy]

[Inaudible]

[Majid Jowhari, Liberal, Richmond Hill, ON]

Okay, I'll come back to you.

[The Chair, Liberal, Dan Ruimy]

Thank you.

Before I move to Mr. Lloyd, I have a question.

[Dane Lloyd, Conservative, Sturgeon River—Parkland, AB]

Is this going to take my time?

[The Chair, Liberal, Dan Ruimy]

No, it's going to take my time. I'll siphon it off somebody else.

Ms. Medeiros, in the beginning of your presentation you had mentioned AI learning, and that in order for it to learn, if I have this right, it would go online and take images. Does that also apply to text?

[Maya Medeiros]

Yes, it can learn from text or video images.

[The Chair, Liberal, Dan Ruimy]

For the purpose of this conversation, could it, then, access Mr. Thompson's book?

[Maya Medeiros]

It could access that book. Whether the trainee of the system or the owner of the system would do that without their permission probably depends on that person. Those who are being diligent and law abiding would not use that without his permission, because there's an uncertainty right now in the law.

[The Chair, Liberal, Dan Ruimy]

So it's programmable. It's not something it would do on its own, go out, like Google, with all the little robots—“Oh, this is a good part of a book, I'll take a chapter out of here” or “I'll take a page out of this book”.

I guess what I'm asking is this. Is this going to present even more of a challenge to writers?

[Maya Medeiros]

The technology can be set up to crawl along, as you mentioned. Some technologies might be doing that already, crawling the Internet, trying to find whatever they can out there and learn from it.

Whether a company is going to take those documents, make copies, bring them in-house, and train on their own systems depends on the lawful acts of that company. A lot of them will probably refrain from doing that, given the uncertainties in the law right now.

Conceivably, from a technology perspective, that kind of system can be created.

[The Chair, Liberal, Dan Ruimy]

Is there any way to track whether that's happening, currently?

[Maya Medeiros]

I don't know of a way to track whether that's happening, but there could be systems that are hosting the book. If a system is offering that book, they might notice that a crawler has come in and accessed that book.

[The Chair, Liberal, Dan Ruimy]

Assuming that Mr. Thompson has a lock on his book, can the AI program circumvent that lock?

[Maya Medeiros]

It conceivably could, if the technology is set up to do that. That would be a violation of the technology protection measure. There are technologies that can circumvent technology protection measures that exist out there. You could combine some technologies.

[The Chair, Liberal, Dan Ruimy]

Thank you.

Mr. Lloyd, it's all yours. You have three minutes left.

[Dane Lloyd, Conservative, Sturgeon River—Parkland, AB]

The chair and I are kind of nerding out on this issue. I appreciate your commentary, Ms. Medeiros.

One interesting thing that was brought up at a previous committee meeting with some indigenous witnesses was that oral tradition and oral knowledge isn't protected under copyright, because one of the tests is fixation. It must be written.

I'm interested to know if audiobooks are protected under copyright.

[Kevin Williams]

Yes, they are.

[Dane Lloyd, Conservative, Sturgeon River—Parkland, AB]

How are they, if they're not fixated?

[Kevin Williams]

It's a performance of the work, and the recording of the performance becomes a fixation of the work.

[Dane Lloyd, Conservative, Sturgeon River—Parkland, AB]

Interesting. Thank you for that, I didn't know. That's good for my knowledge.

This one will be more for Ms. Medeiros again.

I guess I already know the answer to this, but I want to get it on the record. Can artificial intelligence be programmed to deal ethically with copyright?

[Maya Medeiros]

It can be programmed to deal ethically with copyright in the event that there is, let's say, an embedded tag thing—do I have permission to copy this work, or not? It could listen to that tag. It's a computer code that would run that. It could be set up that way.

[Dane Lloyd, Conservative, Sturgeon River—Parkland, AB]

Thank you for that.

Can AI become a tool for monitoring copyright infringement and copyright enforcement?

[Maya Medeiros]

It could in the sense that it can detect patterns and similarities between texts, so it could be used to detect copying, and that would help in the enforcement.