Thank you, Mr. Chair. We appreciate the opportunity to participate in your review.
Google has over 1,000 employees across four offices in Canada, including over 600 engineers working on products used by billions of people worldwide, and ads and cloud teams helping Canadian businesses make the most of digital technology.
Canadian businesses and creators of all kinds use our products and services to connect with consumers and monetize their audiences. According to a recent economic impact study published by Deloitte, which I believe has been distributed to the committee, businesses, publishers and creators generated up to $21 billion in economic activity last year alone, supporting hundreds of thousands of jobs.
The essence of Google’s remuneration and revenue models is a partnership model. Creators such as publishers, producers and developers supply content, while we provide distribution and monetization, technical infrastructure, sales, payment systems, business support and other resources. We then share the resulting revenue, the majority of which goes to the creator every time.
Partnership means that we only earn revenue when our partners earn revenue. It is in our interests to ensure our partners’ success and sustainability. This is why we invest significantly in technology, tools and resources to prevent piracy on our platforms. The Internet has enabled creators to connect, create and distribute their work like never before to build global audiences and sustainable revenue streams, but this new creative economy must ensure that creators can both share their content and make money from it, including cutting off those who would pirate that content.
Five key principles guide our substantial investments in fighting piracy: create more legitimate alternatives; follow the money; be efficient, effective and scalable; guard against abuse; and provide transparency.
The first principle is to create more and better legitimate alternatives. Piracy often arises when it is difficult for consumers to access legitimate content. By developing products that make that easy to do, Google helps to both drive revenue for creative industries and give consumers choice. For instance, the music industry has earned over $6 billion in ad revenue from YouTube, including $1.8 billion in the last year alone.
To do this, we offer a variety of services: ad-supported services like YouTube, subscriptions like Google Play Music and YouTube Premium, and transaction-based services like Google Play Movies & TV. We also support emerging forms of monetization, such as in-app purchases in Google Play Games, and YouTube memberships and Super Chat, which allow users to directly support their favourite creators. Also, we’re finding new ways to allow creators to develop other revenue streams, such as merchandising, ticketing and brand sponsorships.
We want creators to diversify revenue and reduce dependence on ads or subscriptions. This not only helps them build sustainable creative businesses but also insulates them against the negative impacts of piracy.
The second principle is to follow the money. Sites dedicated to online piracy are trying to make money. We need to cut off that supply. Google enforces rigorous policies to prevent these bad actors from exploiting our ads and monetization systems. In 2017 we disapproved more than 10 million ads that we suspected of copyright infringement and removed some 7,000 websites from our AdSense program for copyright violations.
Third is to provide enforcement tools that are efficient, effective, and scalable. In Search, we have streamlined processes to allow rights holders to submit removal notices. Since launching this tool, we’ve removed over three billion infringing URLs. We also factor in the volume of valid removal notices in our ranking of search results.
On YouTube, we’ve invested more than $100 million in Content ID, our industry-leading copyright management system. Content ID allows rights holders to upload reference files and automatically compares those files against every upload on YouTube. When Content ID finds a match, the rights holder can block the video from being viewed, monetize the video by running ads against it or leave the video up and track its viewership statistics.
Over 9,000 partners use Content ID. They choose to monetize over 90% of the claims—and 95% in the case of music—and we’ve paid out over $3 billion to these partners. Content ID is highly effective, managing over 98% of copyright issues on YouTube and 99.5% in the case of sound recordings.
These are just a few of the enforcement tools that we make available for creators and rights holders.
Principles four and five are to guard against abuse and provide transparency. Unfortunately, some do abuse our tools, making false claims in order to remove content they simply don't like. We invest substantial resources to address this and publish information on removal requests in our transparency report.
Google is generating more revenue for creators and rights holders and doing more to fight back against online piracy than it ever has before. Intermediary “safe harbours”, such as the measures clarifying liability of network and hosting services, introduced in 2012, are essential to this.
Indeed, such protections are central to the very operation of the open Internet. If online services are liable for the activities of their users, then open platforms simply cannot function. The risk of liability would severely restrict their ability to allow user content onto their systems.
This would have profound effects on open communication online, severely impacting the emerging class of digital creators who rely on these platforms for their livelihood and curtailing the broad economic benefits that intermediaries generate.
Similarly, limitations and exceptions in the act, such as fair dealing, provide critical balancing by limiting the exclusive rights granted so as to encourage access to copyrighted works and allow for reasonable uses.
One of these uses is information analytics, also referred to as text and data mining. In order for machine learning systems to learn, they need data-based training examples, and it is often necessary for the data sets to be copied, processed and repurposed. In some cases, these data sets may include material protected by copyright, like training an automated text translation system using a corpus of books translated into multiple languages. Unless there is an exception to allow this technical copying, processing and storage, machine learning could infringe copyright, even though the algorithm is merely learning from the data and not interfering with any market for that data or impacting the use by the authors.
It is unclear whether this activity would fall within existing exceptions, putting the Canadian government's substantial investments in artificial intelligence and Canada's significant competitive advantage in this field at risk. We strongly recommend the inclusion of a flexible copyright exception that would permit these types of processes and give much-needed certainty.
I'm happy to discuss these issues with you in more detail and I look forward to your questions.
Thank you.