Canadian Heritage Committee Urges Clear Labeling and Copyright Reform for AI-Generated Content A Canadian House of Commons committee has recommended mandatory labeling for AI-generated images and videos to distinguish them from real content. The report also calls for an expansion of Canadian copyright law to cover AI-generated works and require prior consent for using copyrighted material in AI training data. These recommendations precede the government's forthcoming AI strategy and aim to protect creators' integrity and Canada's digital sovereignty. The rapid advancement and widespread adoption of generative artificial intelligence (AI) tools, exemplified by platforms like OpenAI’s ChatGPT, have led to an alarming increase in hyper-realistic AI-generated images and videos circulating online. Recognizing the potential for misinformation and the erosion of trust, a Canadian House of Commons committee has formally recommended that all AI-generated content, including videos shared on digital platforms, be clearly and unequivocally labeled. This measure is intended to empower individuals to discern between authentic human-created content and sophisticated AI fabrications. In a report formally presented to the House of Commons this week, the Canadian Heritage committee has put forth a series of recommendations aimed at safeguarding both consumers and creators in the evolving digital landscape. Beyond the imperative for clear labeling, the committee proposes a significant broadening of Canada’s copyright law. This expansion would specifically address AI-generated content, ensuring the integrity and protection of creative works. A key element of this proposed reform is the requirement for prior consent before any copyrighted material, encompassing literature, art, and music, can be utilized for training AI models. These proposals are particularly timely, coinciding with the imminent release of the Canadian government's comprehensive AI strategy, spearheaded by Artificial Intelligence Minister Evan Solomon. The committee’s report also champions the cause of digital sovereignty, urging Ottawa to invest strategically in Canadian AI infrastructure – a consideration that Minister Solomon has indicated is actively under review. The concerns voiced by artists are not new; for years, they have implored the government to enact legislation that extends copyright protections to AI-generated content. This includes music crafted to emulate the distinctive style of a songwriter or visual art produced by AI in the likeness of an established artist's unique aesthetic. The existing Copyright Act was fundamentally designed to protect works originating from human creators. The urgency of this matter was underscored at a recent three-day summit on AI and culture held in Banff, Alberta, attended by both Minister Solomon and Canadian Identity Minister Marc Miller. At this event, Canadian artists passionately argued that prioritizing legislation for copyright protection against AI appropriation should be a paramount concern. The current methodology for training AI systems involves the analysis and reproduction of vast datasets, often including copyrighted works, to identify patterns and generate predictions. However, this process has drawn widespread criticism from creators who allege that their work has been incorporated into AI training without their explicit consent or any form of remuneration. While a significant portion of witnesses who testified before the House committee acknowledged the potential of various AI tools to enhance efficiency and foster creativity within cultural sectors, the MPs' report ultimately emphasizes the necessity of regulating the potentially harmful outcomes of AI development to protect Canadian citizens. Some experts who appeared before the committee expressed profound concerns, warning that AI risks crossing a critical threshold, shifting from a tool that serves human creativity to one that entirely displaces it. This sentiment resonates even as experimental platforms, such as OpenAI’s Sora app, demonstrate the capability of generating remarkably lifelike videos, as showcased by The Globe's Samantha Edwards in her illustrative examples. However, a contrasting perspective was offered by Eric Chan, an artist and creator in residence at Library and Archives Canada. He posited that AI represents the current era’s equivalent of the printing press, a transformative technology that, like past advancements in reproduction, was initially met with alarm but eventually became an indispensable part of infrastructure. Chan suggested that AI should not be framed as an existential threat to creators. The prevailing sentiment among most witnesses representing the creative industries was that AI-generated creative output should not automatically qualify for copyright protection unless it demonstrably incorporates a meaningful level of human intervention. The committee was also informed that AI can produce information that is unreliable, misleading, or incomplete. The proposed labeling system, the report suggests, would serve as a crucial tool in clearly distinguishing content, thereby safeguarding the intrinsic value of human creative work and promoting greater transparency. Taylor Owen, the founding director of McGill University’s Centre for Media, Technology and Democracy, voiced strong support for mandatory watermarking of AI-generated videos and images, alongside the requirement for social-media platforms to label AI content. He stressed that knowing the origin of content is essential for Canadians to navigate the increasingly AI-infused information environment, calling it critically important. Furthermore, the House committee’s report advocates for increased transparency from AI developers regarding their use of copyrighted works in training their models. This includes mandating the disclosure of training data sources, which would facilitate the implementation of proper authorization and licensing frameworks. The current situation is further complicated by a growing number of lawsuits filed by creators against technology companies. These legal actions aim to address the unauthorized use of copyrighted works for AI training. Notably, a class-action lawsuit was filed against Google in the United States in 2024, seeking compensation for visual artists and authors whose registered copyrighted works were allegedly used by the tech giant without permission to develop its generative AI models. A similar lawsuit has also been initiated against OpenAI, challenging its practices concerning the use of copyrighted literary, dramatic, musical, and artistic works. Allegations in this suit include the purported violation of copyright law through the scraping of proprietary news content without consent or payment to train its models, such as those underpinning ChatGPT. The lawsuit further includes claims of breach of contract and unjust enrichment against OpenAI, highlighting the complex legal battles emerging at the intersection of AI development and intellectual property rights