Industry Committee on May 11th, 2018

[Majid Jowhari, Liberal, Richmond Hill, ON]

Even if it's just used for making sure that the logic works, once the logic is completed and it's launched, does it continue reaching out to various sources to accumulate more and learn more? There's an initial dataset that is used and then there's an ongoing dataset. Right now, both of those datasets, based on the Copyright Act, are subject to copyright fees.

[Maya Medeiros]

There is uncertainty as to whether the underlying work itself.... If it is a digital copy of a book, it would be a literary work regardless of what stage. Think of natural language processing, for example. You would want to teach a system how to understand a sentence and context and meaning and sentiment. You would probably want to give it a lot of different books and a lot of different texts to figure that out. That would all be part of the training process, so when you fed it something in real time, and not in the training stage, it could figure out the context of that sentence and the meaning of that sentence.

In the training process, it's actually unclear, under the current act, whether that activity would be an infringement of copyright or not. There's uncertainty there, so that often prohibits using works that a company might want to use for that training process. They're not releasing those works, and they're not a substitute for the digital text, but they're using them to learn how those sentences are structured in that process.

[Majid Jowhari, Liberal, Richmond Hill, ON]

That's exactly the point I was trying to make. When it comes to machine learning, you need a large set of data. So whereas when we are using copyright, we're saying let me copy 10% of this one story or this one poem out of a list of all the poems, within machine learning, you have to provide almost a whole book, or a book of poems, or a book of pictures, and all of those. Then no so-called 10% rule applies.

How do we deal with that? It's going to come and it's just around the corner.

[Maya Medeiros]

I'm not sure how you would deal with that. I think there should be some clarity as to whether or not that's permitted under the current framework. I think the 10% rule is a little bit different in this context, because you're not giving 10% of the book to a student or to another person and saying, “Don't buy the book”. It's not a substitute for the original intent of that work. Ideally you want more than just 10 poems. You want every poem in the world if possible.

That's the diversity, particularly looking at translations and what have you. If you're only looking at one poem or a set of poems and saying that this is what is good but all of those poems are created by one type of person, then that notion of what's good or what has a happy sentiment is flawed or biased because you don't have access to a larger training set.

[Majid Jowhari, Liberal, Richmond Hill, ON]

Let's go to blockchain. We know based on the encryption and the distributed model, it's gaining recognition as being able to help us deal with issues such as cybersecurity or even infringement in information.

Given the fact that we need this large base of information or input, how would blockchain be able to help us?

[Maya Medeiros]

I think blockchain would be able to help us in the sense that it can automate decision-making using a smart contract for distributing payments, for example, or for tracking usage. It's just a way to upload that data. It could use a huge dataset. Then you could run artificial intelligence systems over that dataset as another training set if you wanted, as another data structure, and you could understand usage or patterns and that kind of thing from it. I think blockchain is helpful particularly with smart contracts embedded in it for distributing payments or tracking usage in a way such that it doesn't enable one centralized source to handle all of that data.

[Majid Jowhari, Liberal, Richmond Hill, ON]

Okay, great.

[The Chair, Liberal, Dan Ruimy]

That brings to a close the second panel. I really appreciate all of your input. We've been hearing these stories all across the country. There's a lot of work that we have to put together, trying to marry the new and old technology and making sure that culture can thrive.

We are going to adjourn for the day. We will be back at 7 o'clock for the open-mike session. If you want to come back and spend two minutes and get to the heart of the matter and say what you have to say, I encourage you all to come back.

Thank you very much.

[Maya Medeiros]

I have one question about the process. I understand that written submissions are going to be accepted. Is there a timeline around when they should be submitted?

[The Chair, Liberal, Dan Ruimy]

Currently, if you go to the INDU committee's home page, you can find the report, and you will see a button for briefs. You can submit a brief. You can submit correspondence. There is no timetable because this is a longer study.

[Maya Medeiros]

Okay.

[The Chair, Liberal, Dan Ruimy]

We haven't put a time limit on it as of yet.

Is that correct?

[David Groves, Committee Researcher]

If you give it to us before the summer, we'd love it because then we can read it over the summer.

[The Chair, Liberal, Dan Ruimy]

They have nothing to do over the summertime.

Thank you very much.

The meeting is adjourned.