Can web scraping and data harvesting for AI model training survive legal challenges under copyright law?

For AI, tech and Internet business legal services call 403-400-4092 / 905-616-8864 or email Chris@NeufeldLegal.com

Whether web scraping and mass data harvesting for artificial intelligence model training can survive legal challenges under copyright law is one of the most fiercely litigated questions in global intellectual property law. While AI developers maintain that harvesting publicly accessible digital content is a lawful, transformative necessity for machine learning, rights-holders argue that automated data scraping constitutes mass copyright infringement and unauthorized commercialization of proprietary assets. Under Canadian law, resolving this dispute depends on whether data ingestion fits within Canada's strict statutory fair dealing exceptions, how courts interpret contractual website terms of service, and whether AI training outputs substitute for the original works.

Key Legal Insights and Perspectives

1. Canada's Closed Fair Dealing Framework vs. Broad U.S. Fair Use

The legal survival of web scraping in Canada faces a narrower statutory hurdle than in the United States:

  • Statutory Purpose Test: Unlike the open-ended "fair use" doctrine under U.S. copyright law, Section 29 of the Canadian Copyright Act relies on a closed list of statutory fair dealing purposes (research, private study, education, parody, satire, criticism, review, and news reporting). AI developers must first prove that commercial web scraping to build LLMs or foundation models fits within a recognized statutory purpose, such as "research".
  • The Six-Factor CCH Fairness Test: If a statutory purpose is established, Canadian courts evaluate fairness using the six-factor test set out by the Supreme Court of Canada in CCH Canadian Ltd. v. Law Society of Upper Canada: purpose, character, amount, nature, effect on the market, and available alternatives. Commercial AI vendors face significant legal scrutiny under the "effect" factor if scraped training data generates outputs that compete directly with the original content creator.

2. High-Profile Canadian Litigation and Judicial Challenges

Canadian courts are actively grappling with major copyright claims brought by publishers, media organizations, and database owners against AI developers:

  • Media Industry Lawsuits: Major Canadian news publishers (including Toronto Star, Globe and Mail, CBC, and La Presse) have launched significant copyright lawsuits against OpenAI and other AI developers. The plaintiffs argue that harvesting proprietary news archives without licensing agreements constitutes mass copyright infringement, breach of website terms of service, and unjust enrichment.
  • Jurisdictional and Procedural Rulings: In landmark procedural decisions such as Toronto Star Newspapers Limited v. OpenAI, Canadian courts affirmed jurisdiction to hear copyright infringement and data scraping claims against foreign AI vendors, clearing the way for substantive trials on data harvesting legality.

3. Technical Protection Measures (TPMs) and Terms of Service Enforcement

Legal challenges to data scraping extend beyond statutory copyright into contract and digital lock rules:

  • Circumvention of Digital Locks: Under Section 41 of Canada's Copyright Act, circumventing a Technical Protection Measure (TPM), such as paywalls, anti-scraping CAPTCHAs, or IP-blocking controls, is an independent statutory violation. AI scrapers that bypass technical barriers risk statutory liability regardless of fair dealing defenses.
  • Breach of Website Terms of Use: Websites increasingly incorporate explicit "no AI scraping" directives into their terms of service and robots.txt files. Courts are examining whether automated scraping in direct violation of express website terms gives rise to enforceable breach of contract claims.

4. The Shift Toward the "Permission Economy" and Data Licensing

Persistent litigation risks are forcing a structural shift in how commercial AI entities source training data:

  • Commercial Data Licensing: Rather than relying solely on unauthorized web scraping, enterprise AI developers are increasingly entering into multi-million-dollar data licensing deals with digital publishers, media houses, and content platforms to secure legally defensible training pipelines.
  • Legislative Reform Considerations: Government consultations led by Innovation, Science and Economic Development Canada (ISED) continue to examine whether Canada should introduce explicit Text and Data Mining (TDM) statutory exemptions or opt-out regimes to balance technological innovation with rights-holder compensation.

For Artificial Intelligence (AI), technology and Internet business legal matters, contact our law firm to schedule a confidential initial consultation at 403-400-4092 [Calgary, Alberta], 905-616-8864 [Greater Toronto Area] or Chris@NeufeldLegal.com.

Selected Sources & References

AI Law - Introduction