Meera Nair

Posts Tagged ‘B. Zorina Khan’

Looking ahead

In Posts on March 1, 2024 at 7:20 am
an abstract representation of AI and automation, by 紅色死神 CC-BY-NC-SA, Flickr

This past week marked a Canadian festivity known as Fair Dealing Week. One might call it our spinoff of the American tradition of Fair Use Week. In the United States, Fair Use is a provision within their copyright law which limits the control exerted by copyright. Canada’s allowance of Fair Dealing is similar but far more constrained in application. Both limitations draw from principles established by courts and legislators over the past three centuries, wherein some unauthorized and unpaid uses of intellectual creations were deemed essential for the ecosystem of creativity and innovation to thrive.

There is no better exemplar of successfully using copyright limitations for economic development than the United States. In Inventing Ideas (2020), renowned economist B. Zorina Khan examines how a former colony was able to become a leading economy in less than one century, and the world’s technological and industrial leader within two. Copyright strategy was part of the answer; America’s initial objective “was not to benefit authors or publishing companies per se, but to increase contributions to knowledge and the dissemination of information.” Their foundational copyright system promoted the production of inexpensive books and limited the protection thereof. The net result was increased reading, learning, creativity, and innovation.

As I wrote in some detail in 2018, billion-dollar industries consistently emerged in the United States. (Such technological success did not impede the rise of highly acclaimed writers, musicians, artists, and other creators.) Copyright policy only shifted in priority after achieving international dominance; in the late twentieth century, control became the central goal. American heft in the global market meant that other nations could be either cajoled or threatened into following suit, even though such measures prevented those countries from emulating America’s development path.

Yet it remains that the grant of copyright is not all-powerful. In fact, by virtue of the Berne Convention, all participating countries must provide some allowance of unauthorized use; Global Mandatory Fair Use by Tanya Aplin and Lionel Bentley comes to mind. For Canadians, Fair Dealing provides a modest degree of copying of protected content to support learning, research, journalism, and some creative undertakings. But that will not be enough if Canada is serious about Artificial Intelligence development.

Again, the United States is the illustrative example. Fair Use paved the way for using existing content to develop the large language models that underpin AI programs. Lawsuits have abounded over the training of those models, with content owners arguing they should be compensated when their work was used without consent, yet it is not at all certain that they will win.

But winning may not be necessary. Dominant companies in the AI sector are increasingly offering compensation to larger media organizations for the use of their content as training data. Even though, as prominent legal scholar Andres Guadamuz writes, “training an AI is not an exclusive right of the author.” While some will be pleased at the undercurrent of willingness to pay by well-heeled tech companies, we should be wary. If that practice becomes the norm, who can enter and compete in that field? Will basic research even be possible? It should not come as a surprise that the likes of Alphabet, Meta, and Microsoft are willing to pay, as it will help ensure their continued dominance in this sector.

It is disappointing that the present Canadian government routinely rails against such companies, but is unwilling to take the steps to allow our domestic AI talent to thrive and compete in the global tech-sector. We are falling behind our competitors.

For instance, Israel has provided guidance to confirm that their regime of Fair Use could serve AI-development in a similar fashion to what American companies enjoy. Asia Pacific countries have, and are, moving ahead with AI-friendly copyright regimes; Peter Yu (another highly respected legal scholar) draws attention to the systems already in place in Japan and Singapore, and adds that “[Chinese copyright laws may be] very supportive of AI development once the appropriate regulations have been introduced.” Even the old guard, the United Kingdom, provided some support for the analyses that underwrite AI; although now the UK expressly requires licenses for commercial activity.

Whereas Canada seems content to continue our tradition of producing a highly skilled workforce to support foreign corporate owners who either set up branch operations here or just cut-to-the-chase and lure that expertise southward. Last year, for the Senate Standing Committee for Banking Commerce and the Economy, Jim Balsillie expressed great concern at the “exfiltration of knowledge assets out of Canada.” Among his examples, was this:

Foundational IP for AI that Canadian taxpayers funded for two decades is transferred from the University of Toronto to Google, who thanked Canada for it and said: “We now use it throughout our entire business and it’s a major driver of our corporate success.”

If Canada wishes to compete in the AI space, it has much catching-up to do. A place to start is implementation of recommendations provided in 2019 by a credible Standing Committee; their considered opinion was that Canada should expand the limitation of Fair Dealing to resemble something more akin to Fair Use and facilitate informational analysis as would support AI development.

Then we could really celebrate the importance of limitations within systems of copyright.