Artificial Intelligence Information

• Information
  
  BCFSA’s information is clear, concise, easy-to-read explanations of the requirements for real estate professionals under the Real Estate Services Act (“RESA”), Real Estate Services Regulation (“Regulation”), Real Estate Services Rules (“Rules”), and other applicable legislation.

  This information is intended for use by real estate professionals, to support their understanding of the standards they must meet in the delivery of real estate services.

What Is AI?

While there are many definitions, artificial intelligence (“AI”) generally refers to the use of computers or computer-based systems to perform tasks commonly associated with intelligent beings.¹ Examples of “intelligent processes” performed by AI include learning,^2,3 reasoning,^4,5 generalizing,⁶ inferring meaning,⁷ and creativity⁸ (e.g., generation of new content). By comparison, “generative AI” specifically refers to a subset of AI focused on generating new, original content. Generative AI systems are designed to create data, images, text, or other types of content that are not explicitly programmed.

Many forms of AI exist, ranging from rules-based systems designed to automate decision-making within defined parameters⁹ (e.g., processing routine claims or applications) to more complex machine learning systems that perform prediction, classification, and decision-making tasks based on analysis of large data sets.¹⁰

AI Benefits and Risks

The strengths of AI include the ability to automate routine processes, allowing people to focus their time and efforts on more complex or unique situations. AI systems can also analyze large data sets and draw insights or conclusions, often faster than a person could.

However, there are a number of important critiques to consider, including the potential for AI to reproduce biases found in user input or training data as well as other errors in decision-making (e.g., false matches of facial recognition technology for visible minorities and women¹¹). From a privacy perspective, it is difficult for consumers to fully understand how their data and information will be used by AI, resulting in challenges related to:

• Lack of explicit consent;¹²
• Potential for unintended uses;¹³ and
• Potential to cause harm to individuals and/or their interests.¹⁴

Legislative AI Framework in Canada and B.C.

In September 2023, the federal government launched a voluntary code of conduct on the development and management of generative AI systems. However, there is currently no specific statute that regulates the development or use of AI in Canada or B.C. As of January 2024, the federal government is considering proposed legislation called the Artificial Intelligence and Data Act. If passed, it would regulate private sector organizations that design, develop, or make available for use artificial intelligence systems.

The utilization of AI is presently governed by a broader legislative landscape, encompassing elements of privacy legislation, human rights legislation, and tort law. It is important to note that this landscape is fragmented, with significant gaps across these different pieces of legislation.

Application of Licensee Duties to Clients in Relation to AI

Many AI applications are available to businesses and the public, including real estate licensees. Licensees’ duties to clients provide parameters and guidance to licensees when considering the implications of using these systems in their day-to-day real estate practice. Moreover, licensees’ duties to clients are a reminder that licensees remain accountable for their actions, despite the usage of AI tools, and that they will be disciplined if their actions do not comply with the regulatory framework. Examples of how licensees’ duties to clients may apply to the use of AI are provided below:

• Duty of confidentiality – many AI systems use the information gathered from other users through previous queries to build and train the dataset used to respond to subsequent inquiries. Information about clients and their interests (e.g., names and personal identifiers, proprietary information), even when anonymized,¹⁵ may be unintentionally exposed to others if entered into an AI system.

• Duty to act in client’s best interest – similar to the duty of confidentiality, the potential to expose a client’s information or interests through an AI system can create risk of harm to the client’s interests or position.
• Duty to disclose material information – licensees have a duty to disclose material information to their client, including all material information related to the use of their information in an AI tool. In order to provide meaningful consent to provide their information to AI systems,¹⁶ clients need to understand what information is being collected and how, and to what purpose the information will be used. With AI systems, the potential for unintended or unanticipated uses of information makes it difficult for clients to fully consent to provide their information to AI.¹⁷

• Duty to act with reasonable care and skill – licensees have a duty to ensure the accuracy of any representations they make or information they give that their client may rely on for decision-making.

AI’s Impact on Copyright and Intellectual Property

AI-generated content may not be copyright protectable under the current Canadian Copyright Act. However, it is anticipated that this landscape will evolve as other jurisdictions are progressively recognizing ownership rights for AI-generated work under specific circumstances. That said, licensees aiming to publish works they intend to safeguard might consider refraining from using AI tools, given the prevailing likelihood that such content may presently reside in the public domain if devoid of human contribution.

Since AI may be trained using copyrighted works or associated metadata, it is possible for users to direct an AI to mimic or copy an original work or style. However, most companies providing AI systems have terms of service that prohibit users from intentionally using prompts that draw on copyrighted works to generate an output. Consequently, licensees are advised against instructing AI tools to employ copyrighted content for generating outputs, as this could contravene the AI company’s terms of service and potentially result in legal repercussions from the copyright holder. Should licensees intend to employ specific copyrighted works, securing written permission and retaining documentation is imperative.

Additionally, if AI-generated content incorporates third-party data or information, licensees must ensure diligent adherence to proper attribution and compliance with licensing terms.

1. ^{Artificial intelligence (AI) | Definition, Examples, Types, Applications, Companies, & Facts | Britannica, accessed on July 5, 2023.}
2. ^Ibid
3. ^{What is artificial intelligence (AI)? — AI definition and how it works (techtarget.com), accessed on July 5, 2023.}
4. ^{Artificial intelligence (AI) | Definition, Examples, Types, Applications, Companies, & Facts | Britannica, accessed on July 5, 2023.}
5. ^{What is artificial intelligence (AI)? — AI definition and how it works (techtarget.com), accessed on July 5, 2023.}
6. ^{Artificial intelligence (AI) | Definition, Examples, Types, Applications, Companies, & Facts | Britannica, accessed on July 5, 2023.}
7. ^{Artificial intelligence (AI) | Definition, Examples, Types, Applications, Companies, & Facts | Britannica, accessed on July 5, 2023.}
8. ^{What is artificial intelligence (AI)? — AI definition and how it works (techtarget.com), accessed on July 5, 2023.}
9. ^{3546 (oipc.bc.ca), pg. 7.}
10. ^{3546 (oipc.bc.ca), pg. 8.}
11. ^{3785 (oipc.bc.ca), pg. 18}
12. ^{3785 (oipc.bc.ca), pg. 15}
13. ^{3546 (oipc.bc.ca), pg. 14}
14. ^{The Artificial Intelligence and Data Act (AIDA) – Companion document (canada.ca), accessed on June 21, 2023}
15. ^{3546 (oipc.bc.ca), pg. 41}
16. ^{3546 (oipc.bc.ca), pg. 24}
17. ^{3546 (oipc.bc.ca), pg. 21}