Copyright in the Age of AI: What Law Firms Need to Consider

Artificial intelligence (AI) is no longer an idea we see played out in science fiction movies. From predictive analytics to AI-assisted drafting, machine learning is reshaping the way law firms operate. But as AI becomes more embedded in legal workflows, Copyright Law has entered a gray zone.

For law firms in the United States and Canada, the intersection of Copyright Law and Artificial Intelligence raises complex questions with direct implications for client counseling, risk management, and even the practice of law itself.

The Copyright Puzzle: Who Owns AI-Created Works?

At the heart of the issue is authorship. Copyright traditionally requires a human author. Both U.S. and Canadian Copyright regimes are premised on the idea of creative expression by natural persons.

• In the U.S.: The Copyright Office has made clear that works “not created by a human being” are not protected. In 2023, the Office reiterated this in Zarya of the Dawn, refusing copyright registration for AI-generated images created with Midjourney, though it allowed protection for the human-authored text accompanying them. If interested, you can read the Copyright Office’s letter response in connection with this matter here: https://www.copyright.gov/docs/zarya-of-the-dawn.pdf

• In Canada: The Copyright Act does not explicitly define “author” as human, but courts and the Copyright Board have historically linked authorship to human creativity. While Parliament has not yet legislated specifically on AI, the Copyright Office of Canada is actively studying the issue, and reform is expected. You can read an informative consultation by the Canadian government on the topic here: https://ised-isde.canada.ca/site/strategic-policy-sector/sites/default/files/attachments/2022/ConsultationPaperAIEN.pdf

For law firms, this means that clients relying on AI to generate contracts, marketing content, or creative assets may not be able to claim copyright protection. This is inevitably an important point to consider when advising on risk and ownership.

Training Data and Infringement Risks

Another pressing issue is how AI models are trained. Most generative AI systems are built on vast datasets scraped from publicly available works, many of which are protected by Copyright. This raises legal questions, such as:

• Does using copyrighted works to train AI systems constitute a form of infringement?

• Does it fall under “fair use” (U.S.) or “fair dealing” (Canada)?

  
  

The answers are unsettled. In the U.S., multiple lawsuits are testing these boundaries, including high-profile claims against AI companies for copying millions of works without permission. In Canada, fair dealing is narrower than U.S. fair use, which means companies deploying AI tools may face greater legal exposure north of the border.

For law firms, this creates two obligations: to monitor ongoing litigation closely and to help clients understand the risks of adopting AI systems without clear licensing arrangements.

Law Firms as AI Users: Managing Input and Output Risks

Beyond advising clients, law firms themselves are AI users. Tools that summarize caselaw, draft memos, or assist with discovery can be powerful efficiency boosters. But they also raise two key copyright-related issues:

1. Input Risks

   • Uploading client documents into AI platforms can expose confidential and copyrighted information to third-party vendors.

2. Output Risks

   • AI-generated content may not be copyrightable and could potentially infringe on existing works.

     
     

To minimize risk, law firms should review vendor agreements carefully, clarifying who owns the AI-generated output and whether the provider indemnifies against infringement claims.

U.S. vs. Canadian Copyright Law: Key Differences

While both countries are grappling with the same fundamental questions, there are notable differences:

• Fair Use vs. Fair Dealing: U.S. fair use is flexible and open-ended, allowing broader justifications for AI training. Canadian fair dealing is more rigid, limited to specific purposes (e.g., research, criticism, private study). This makes AI training potentially more legally vulnerable in Canada.

• Moral Rights: Canadian law gives authors moral rights (the right to attribution and integrity of the work), which may complicate the use of AI outputs that remix copyrighted content.

• Legislative Movement: The U.S. Copyright Office is actively engaging in public consultations and issuing guidance. In Canada, reforms are slower, but they are expected, especially as the government reviews intellectual property frameworks in light of AI.

  
  

Practical Steps for Law Firms

So, what should law firms be doing now? A proactive approach is key. Here are some practical considerations:

1. Audit AI Use: Understand how AI tools are being used in-house and by clients. Are they generating content, analyzing data, or training models?

2. Review Contracts: Ensure client agreements and vendor contracts clearly address ownership, licensing, and liability for AI-generated works.

3. Advise on Risk: Educate clients about the uncertainty around Copyright protection and infringement risks, especially if they are investing in AI-generated content or deploying AI in sensitive areas.

4. Stay Informed: Follow ongoing litigation and regulatory developments in both the U.S. and Canada. Case law is moving quickly, and today’s uncertainty may solidify into tomorrow’s precedent.

5. Develop Policy and Standard Operating Procedures: Establish firmwide guidelines for AI use that balance innovation with professional responsibility, client confidentiality, and intellectual property law. Creating and maintaining procedures helps your staff know how to handle potential risks and use AI tools responsibly. This will not only prepare your team for the rapid changes happening in the AI space, but also ensure that your firm can adapt smoothly and confidently.

   
   

Practical Takeaways for Law Clerks, Paralegals, and Legal Assistants

AI isn’t just a tool for lawyers. Law clerks, paralegals, and legal assistants are increasingly engaging with AI in day-to-day workflows, whether through document review platforms, research tools, or drafting aids. Understanding the Copyright context helps them support the firm’s compliance and efficiency. Some key tips to consider include:

• Flag AI-Generated Content: When drafting templates, correspondence, or research summaries with AI tools, always note which portions were machine-generated to ensure appropriate review by supervising lawyers. Keep in mind that while AI tools are great as tools to assist you, they’re not always accurate and that’s where that extra ‘human eye’ will come in handy.

• Protect Client Materials: Never upload confidential or copyrighted documents into public AI platforms unless the firm has approved protocols in place. Treat AI input the same way you would treat disclosure in discovery.

• Check Vendor Terms: When assisting with onboarding or managing third-party platforms, review the fine print—particularly clauses on ownership of AI outputs and indemnification. This will help you in raising red flags for consideration to the supervising lawyer.

• Stay Organized: Develop filing systems that distinguish between human-created and AI-assisted work product. This helps the firm track potential Copyright concerns down the line.

• Continuous Learning: Keep up with internal training and industry updates. As Copyright Law around AI evolves, clerks/paralegals and/or legal assistants who understand the tools being used (and the issues that may arise from them) will be invaluable resources to their teams.

  
  

By adopting these practices, support staff not only safeguard the firm’s intellectual property compliance but also strengthen their own professional expertise in a rapidly changing legal landscape.

Looking Ahead

The rise of AI represents both a challenge and an opportunity for the legal profession. Copyright Law, with its deep reliance on human creativity, is being stretched to its limits. For law firms, this moment calls for vigilance, ongoing education, adaptability, and leadership.

By staying ahead of developments and offering informed counsel, firms can help clients navigate the uncertainty, while at the same time, modelling responsible, ethical AI use within their own practices.

AI may be rewriting the rules of Copyright, but the legal industry has an essential role in shaping how those rules are applied.