DSNews delivers stories, ideas, links, companies, people, events, and videos impacting the mortgage default servicing industry.
Issue link: http://digital.dsnews.com/i/1511756
MortgagePoint » Your Trusted Source for Mortgage Banking and Servicing News 50 November 2023 F E A T U R E S T O R Y to train them. If you input faulty, outdated, or incomplete data, the output will also be incorrect. Additionally, generative AI can be challenging to implement for banks and lending institutions that still rely on legacy systems and complex IT infra- structures, which may not be compatible with generative AI solutions. On the other hand, businesses that already have adopted automation and have an updated infrastructure in place have the flexibility and agility needed to train, implement, and operate generative AI models. Q: How can mortgage companies prepare for the costs associated with generative AI? Adopting generative AI requires care- ful planning and consideration. Mortgage companies should start by conducting a comprehensive cost-benefit analysis to measure the potential return-on-in- vestment (ROI) and identify areas where implementing generative AI will have a positive impact. The next step should be to develop a budget that covers the implementation costs. Because these are not simple tasks, mortgage companies might consider collaborating with external partners or vendors that specialize in gen- erative AI, which can help with the costs associated with infrastructure, software, and expertise. Q: What are the potential risks associated with generative AI adoption, such as lending bias and security? How can these risks be mitigated? Businesses must have a clear and actionable framework for using generative AI and a plan to be prepared for AI-related risks and mitigate them. Data security, wrong analyses, and lending bias are some of the common concerns that surround generative AI adoption in the mortgage industry. Let us take the risk of lending bias as an example. Again, generative AI models are as good as the data they are trained on. If you feed generative AI models inaccu- rate and biased data, they will amplify that bias and generate a discriminatory result. To address this issue, generative AI models must be trained with diverse and representative data. The training data should appropriately reflect real-world use cases and promote fairness and inclusivity to minimize bias. Additionally, continuous auditing, monitoring, and training of gen- erative AI models can help improve their ability to identify and mitigate any biases that emerge over time. As far as security—which is para- mount in our industry—robust data pro- tection measures like encryption, access controls, and secure storage facilities must be in place to safeguard sensitive informa- tion. Regular security audits of generative AI models can also help identify vulner- abilities and stay current with the best data security practices. Creating policies and guidelines for the ethical usage of AI models can also mitigate security risks. Employees using AI systems must be edu- cated about these risks, so they can use the system wisely and protect sensitive data. Q: How can generative AI and augmented reality be combined to provide predictive insights into things like real estate construction, valuation, and property preservation? What impact will this have on the housing industry? Generative AI and augmented reality (AR) are emerging technologies that can be transformative for the housing industry. When combined, they can fuel a kind of interactive design approach in which data constantly inspires new models and virtual projections that are visually presented. These immersive experiences have enor- mous potential in the future to provide predictive insights. For example, generative AI and AR can be combined to generate realistic 3D models of buildings and enable users to visualize and explore virtual construc- tion sites for a wide range of purposes. They can be used to see how a building's position and design would affect exposure to the sun and local weather patterns, or to appraise properties by analyzing various factors such as location, features, market trends, and historical data to generate predictive valuation models. They can also create immersive virtual property tours, where buyers can visualize potential mod- ifications without being on site. They can help find and correct existing issues in a building's structure, which could also help in property preservation needs, or even predict potential risks and hazards associ- ated with real estate projects by analyzing construction and safety data. Q: How can mortgage companies build a culture of innovation that supports the adoption of generative AI and other emerging technologies? Building a culture of innovation that supports the adoption of disruptive technologies is like unlocking a hidden treasure. It begins with the unwavering commitment of a company's leadership, whose vision sets the stage for a thrilling journey ahead. Companies must also develop an innovation strategy that aligns with the organization's goals and vision. It is also wise to build cloud capabilities so that the company is ready at the time new technologies are implemented. Finally, it should also include training programs for employees to build their knowledge and skills in emerging technologies. Encouraging continuous learning, taking calculated risks, and learning from failures are all part of building a culture of innovation. Such an effort is rarely accom- plished alone, however. When it comes to generative AI and other new technologies, there is so much to learn and so many use cases yet to be discovered. For many lend- ers and servicers, tapping the expertise and resources of a partner that already has a culture of innovation can shine a bright light on the path ahead and guide them on their journey.