Blog

RICS Updates: Using AI in Valuations

The RICS recently updated the Red Book to address the use of AI and automated tools in property valuations. In essence, it states that valuations must combine human expertise with technology to comply with the standards. Here’s what you need to know about the recent proposed updates, with my own thoughts and comments.

Alex Kountourides, Co-Founder, Valos.

1.4 IVS 105 includes the following ‘No model without the valuer applying professional judgement, for example an automated valuation model (AVM), can produce an IVS compliant valuation’. For the avoidance of doubt, the provision of a valuation wholly or partly based on the derived output of one or more of:

  • an automated valuation model (AVM)
  • a valuation modelling tool or valuation software
  • a valuation process software tool or template
  • a data scraping tool or other form of automated data management,

is regarded as the provision of a written valuation for the purpose of these standards (see VPS 5 paragraph 3) subject to the application of professional judgement by a member, which must be done in accordance with the mandatory requirements of these standards

We strongly support the IVS position and its inclusion within the Red Book. The distinction between pure AVM-derived outputs and valuations that combine professional knowledge with technology is crucial. This stipulation safeguards the integrity of Red Book compliant valuation reports, protecting both the profession and public interest.

Automated Valuation Models (AVMs) and other automated tools can indeed aid the valuation process by handling repetitive and administrative tasks. However, these tools cannot account for the breadth of factors that a skilled valuer considers, nor can they substitute the professional judgment of a qualified surveyor which is essential for accurate valuations.

  • We suggest providing more clarity on what constitutes “derived outputs,” as some terms used are broad or potentially misaligned with the intent of the standard.
  • “Valuation software” is an expansive term. Clarification is needed on whether this includes general-purpose software used in valuations (e.g., Word, Excel) or software for specific tasks like comparable evidence searches or map production.
  • The term “template” in the third bullet point isn’t inherently a technological application for output. Consider revising to “template generator,” “document automation,” or “template automation” for greater specificity.
  • The fourth bullet point conflates data scraping (a form of data gathering) with data management. We suggest distinguishing between data-gathering methods (including APIs and downloads) and data management. Additionally, the terminology referring to “output” does not relate to data gathering. Data gathering is an input. We believe that equal attention should apply to both inputs and outputs as understanding inputs is crucial for valuers to pass judgment on the output and therefore assess the applicability and relevance of technology in the valuation process as these inputs ultimately influence the final outputs.
  • A clear definition of “automated data management” would be beneficial.

1.5 The use of Artificial Intelligence (AI) is not prohibited subject to appropriate consideration and documentation of matters relating to intellectual property, data and input/output verification, transparency with the intended user(s) of the valuation and all other ethical, technical and legal matters referred to in these standards.

We commend the RICS’s openness to AI use in valuations. However, this inclusion may raise compliance questions for practitioners and firms. We recommend that RICS provides more detailed guidance on:

  • Specific ways to integrate AI and automated tools into the valuation process while maintaining compliance with standards.
  • Best practices for documenting AI and automated tool usage to ensure transparency and auditability.
  • Guidelines for verifying the accuracy and reliability of data inputs and outputs when using AI and automated tools.
  • Addressing ethical considerations and potential biases in AI systems used in valuations.

Clear guidance in these areas will facilitate the responsible adoption of beneficial technologies while upholding high professional standards.

We strongly endorse the human/AI hybrid model of valuations which we believe provides the best of worlds, enabling the industry (and society) to reap the benefits of technological advancements whilst limiting the potential harms.

 

Stay on top of the latest industry developments – book a demo with Valos today.