Ethical Guidelines for Property Professionals Using Generative AI

Executive Summary

Generative AI is artificial intelligence that creates new content, including text, images, music, video, software code, and designs autonomously. (ChatGPT, Gemini AI etc)

AI in general and Generative AI in particular offers diverse applications in administrative tasks and scheme management services across all agency sizes and practice areas. However, its use must align with professional responsibility obligations. Managing agents need to understand the risks and benefits associated with Generative AI in providing scheme management services.

In addition, when deciding to use Generative AI, specific factors such as the unique needs and challenges of each client, body corporate, scheme, and agency size must also be considered. This document endeavours to provide practical guidance to ensure that Generative AI usage remains consistent with professional responsibilities, emphasising independent critical thinking, integrity, data privacy, and risk management.

Several recent court cases as well as the government’s regulations on the use and administration of AI demonstrated the potential wrong use Generative AI.

1. Integrity and Diligence

• Ultimate Responsibility: Be responsible for the content of your work and aware of Generative AI’s potential to generate incorrect or fabricated information.

• Understanding the Tool: It is the managing agent's responsibility to ensure that he or she understands how the technology works, including its limitations and terms of usage.

• Rigorous Verification: Treat Generative AI as a tool akin to an assistant, with all outputs rigorously checked for accuracy. Validate and correct both the input and output of Generative AI to ensure accuracy and alignment with client and body corporate interests.

• No Delegation of Accountability: In the final analysis, the managing agent is responsible for the professional management services and advice that he or she provides.

  
  

2. Appropriate Use of Generative AI

Given the current state of technology, Generative AI is useful for routine tasks such as:

• Summarising texts

• Improving structure and language

• Transcribing interviews or meetings

However, managing agents are cautioned against overreliance on AI tools and are called to respect the necessity of human judgment.

Using Generative AI to provide information about South African law (such as sectional title regulations) or for more complex tasks such as applying regulations to facts may be helpful but is not currently recommended. If done, it should be done with caution. A managing agent should always draw his or her own conclusions and recognise the limitations of Generative AI.

3. Independent Critical Thinking

A managing agent should maintain an unbiased search for truth and apply rigorous independent thinking. Be aware that Generative AI may prioritise highly cited cases or documents, overlooking relevant but less known authorities or local regulations. Generative AI may also be incorrect and can even refer to non-existent rules or legal precedents.

Conduct thorough searches and independent analyses, proactively seeking out and citing relevant authorities and actual rules. It is worth repeating that the managing agent is responsible for the scheme management services and advice that he or she provides.

4. Quality of input

 One of the most important realities to understand is that Generative AI responds exactly to what it is given. It does not interpret intent in the way a human would. It does not fill in gaps or assumptions unless prompted to do so.

Effective use of Generative AI starts with clear thinking. Before asking a question, the accountant must be clear on what is required, why it is required, and how the answer will be used.

5. Benefit to Client

Generative AI can significantly enhance the efficiency and effectiveness of scheme management services, providing substantial benefits to bodies corporate. Although there is currently no obligation for a managing agent to disclose the use of Generative AI tools to clients, managing agents may exercise their discretion about disclosing AI use to clients or trustees. Generative AI is considered a tool akin to other technological aids in property management practice.

Practitioners should focus on delivering competent and diligent services, with Generative AI enhancing their ability to do so. The use of Generative AI can lead to quicker turnaround times for routine tasks, allowing more time to focus on complex scheme management issues. Additionally, Generative AI can help reduce costs, and these savings should be passed on to clients where appropriate, reflecting the increased efficiency gained through the use of this technology.

Responsible partners and directors of management agencies should issue guidelines on evaluating when to disclose AI use to clients and how to communicate the benefits and risks.

6. Confidentiality

• Terms Review: Review the Terms of Use of Generative AI products to understand how inputted information is utilised and ensure it is not shared with third parties or used to train the Generative AI without consent.

• Protecting Client Data: Avoid inputting confidential client or body corporate information into Generative AI solutions were doing so would compromise confidentiality or professional privilege.

• Security & Anonymisation: Where use would not compromise confidentiality, it is advisable to consult IT professionals to ensure Generative AI systems adhere to stringent security protocols and to anonymise client and scheme information to minimize risks of identification.

7. Bias

Be aware of potential biases in Generative AI outputs. Generative AI is often designed with foreign values in mind and using foreign datasets, privileging foreign issues and perspectives. This can result in outputs that reflect these biases, side-lining local contexts and South African statutory or legal values.

Generative AI is also often designed using statistical or probabilistic models with uncertain reasoning methods. Continuous learning about Generative AI biases and their implications is essential. Managing agents should critically assess Generative AI-generated responses for bias, seek out local regulations, and ensure that their advice is aligned with South African legal rules and property sector values.

8. Cybersecurity and Risk Management

Implement robust security protocols to protect scheme and body corporate data from breaches. Be aware of cybersecurity risks such as prompt injections and data source corruption. Regularly review and update cybersecurity measures. Maintain effective professional quality control over Generative AI output and use.

9. Continuous Learning and Adaptation

• Regularly update knowledge on new developments in Generative AI technology to remain competent in your role.

• Participate in workshops and training sessions to share experiences and insights on the use of Generative AI in property management practice.

• In bigger property management agencies, the agency partners or executives must issue explicit guidelines for managerial and supervisory roles, ensuring they establish policies and provide training on AI use.

Conclusion

The use of Generative AI in scheme management practice offers significant benefits but also poses substantial risks. By adhering to these guidelines, managing agents can leverage Generative AI responsibly and ethically, ensuring compliance with their professional obligations while maintaining the integrity and quality of their service to bodies corporate.

Acknowledgements:

This article was generated with the generative drafting assistance of Gemini AI and

https://law.ukzn.ac.za/ukzn-law-academics-draft-ai-ethics-guidelines-for-legal-practitioners