When ChatGPT came to market last year, it marked a significant advancement in how Generative AI could be used. Recent advancements in artificial intelligence (AI) have enabled companies of all types to deploy more powerful AI models in their operations. At Qualia, we see substantial potential for AI to improve the experience for our users and continue to transform real estate transactions. This new technology presents a host of exciting possibilities. As with the introduction of any new technology though, it also opens up questions regarding how it can most securely and effectively be used.
At Qualia, we are committed to exploring the possible benefits of adopting Generative AI, and we are committed to doing so in a manner that prioritizes security and accuracy. Given how quickly this new type of technology is developing, we believed it was imperative to develop a set of internal guidelines for how we’ll identify, develop, and deploy new features. Our Product, Engineering and Information Security (InfoSec) team outlined a set of operational principles to guide this work. This document outlines those principles. We will operate under those when implementing Generative AI-powered features to ensure our application is done in a secure and efficient way for our users.
A Brief Background on Generative AI
Generative AI, also known as Gen AI, is a type of machine learning that can learn from existing materials, such as text and images. It does this by focusing on patterns and relationships in the materials upon which it is trained. It can also generate new, realistic materials that reflect, but do not directly replicate, the characteristics of the training material.
These new materials are generated in response to one or more questions or prompts from a user. Many Generative AI models rely on large language models (LLMs) to process, generate, and understand vast amounts of data in order to deliver more human-like responses to queries. An LLM is a type of artificial intelligence trained on vast amounts of text to generate and comprehend human-like language based on learned patterns.
Our Goals and Principles
In light of the potential benefits Generative AI can have on productivity, we determined it was necessary to explore how its application could best serve our customers. Our primary goal when considering potential applications of Generative AI within our suite of products is to identify implementations that will empower our users to be more efficient.
As an example, Generative AI can be used to extract unstructured data from a document and convert it into structured data. By converting information into structured data it can then be automatically pulled into an open order, and new features can be developed to provide recommendations for smart actions and workflows. This application of the technology dramatically reduces time spent manually rekeying information into the system and reduces the risk of human error in the process.
While embedding Generative AI in select products offers clear benefits to our users, it also presents new challenges and potential risks that we will carefully consider and manage as this technology develops. Further, it’s important for us, and any technology provider, to understand how integration partners throughout the title & escrow technology ecosystem are managing their use of Generative AI. To that end, we have updated our existing risk management protocols for technologies and vendors to account for risks introduced by Generative AI. We developed our approach to Generative AI usage by consulting with various expert resources, including those from regulators, law firms, think tanks, and government entities like the White House, which published the Blueprint for an AI Bill of Rights.
Below are the principles we have outlined to guide our approach to implementing Generative AI:
- AI must be built with a priority on security first and protection of customer data. We updated our existing data protection and product procurement procedures to take into account the use of Generative AI in the products we create and the systems we integrate with. This includes:
- Protection of customer data: Qualia was built on a foundation of protecting the details associated with Customer Order Data that is processed within a customer’s deployment within Qualia. That includes the personal information about the consumers involved in a residential real estate transaction, such as buyer and seller information (e.g., names, SSNs, etc.). In order to maintain those protections, Qualia acquires enterprise-grade licenses from Generative AI providers to guarantee that user prompts and customer data are not utilized to train their foundational models.
- Technical security reviews: Our dedicated InfoSec team collaborates closely with our Engineering and IT teams to uphold stringent data security standards and obtain third-party certifications for our products. This includes rigorous reviews of our Generative AI provider’s security measures so we continue to protect our customers’ data from potential security and privacy threats.
- Legal safeguards: We will ensure that the providers of Generative AI models, or the applicable cloud service provider (e.g., Amazon Web Services, Microsoft), are contractually obligated to limit their usage of Qualia’s data for specific and limited purposes, protect the data with appropriate measures, and retain the data for a short and predefined period of time.
- AI does not replace humans, it assists them. We believe AI assists humans with their work, it does not remove them from it. Our applications will be designed to keep humans in the loop. AI can be great for automating tasks and providing a summary or interpretation of content, but humans should review the work performed and the content generated.
- AI is to be used when it’s the best option to improve our customers’ efficiency and growth goals, building AI is not the goal itself. We will use AI when it’s the most helpful solution for our users and will constantly evaluate it for effectiveness. Any Generative AI model that we use in our products will be subject to ongoing evaluation once deployed.
- AI will transform the way business is done, but it will not happen overnight. As one of many companies developing features that leverage AI, we understand the rapid pace at which this technology is evolving. It’s essential to approach our development with patience, intentionality, and adaptability, recognizing that the landscape of AI will constantly change and grow. Our commitment is to stay at the forefront of these changes, ensuring our tools are innovative, secure, and reliable.
Our Approach to Innovation
We believe that Generative AI will transform how people work everywhere, including in the real estate industry. And while the potential is great, we are taking a measured and cautious approach to our deployment of Generative AI. Our goal is to deliver the most effective applications of this technology within a real estate transaction while ensuring the safe and secure management of all customer data within our products. We are excited about the potential application of Generative AI, and we look forward to navigating the future of real estate transactions together.