A Phil Hall Op-Ed: Warren Buffett hosted his annual Berkshire Hathaway (NYSE:BRK.A) shareholders meeting this past weekend, and during the event the billionaire executive was quizzed about artificial intelligence (AI).
“I don’t know anything about AI,” he said, according to CNBC. “That doesn’t mean I deny its existence or importance or anything of the sort.”
His answer was both refreshing and disturbing for its honesty. It was refreshing because it offered a rare opportunity for one of the most respected investors and corporate leaders of our time to acknowledge that he was not up to speed on a technology that could change how business functions – to his credit, he wasn’t trying to bluff his way out of the conversation. But it was also disturbing because so many of Buffett’s businesses – including his companies focused on housing and real estate – are figuring out ways to incorporate AI into their operations, and at the very least he should be up-to-speed on how AI will impact his corporate empire.
Buffett sought to offer a historical analogy to the challenges of AI and came up with an example that might be a tad melodramatic.
“We let the genie out of the bottle when we developed nuclear weapons and that genie has been doing some terrible things lately, and the power of that genie is what scares the hell out of me,” he continued. “I don’t know any way to get the genie back in the bottle, and AI is somewhat similar. It’s part of the way out of the bottle, and it’s enormously important and it’s going to be done by somebody. Whether it’s going to change the future of society, we will find out later.”
Actually, Buffett has it wrong on that last point – we need to find out sooner than later. Earlier this year, the Bipartisan Policy Center published an interesting study on how AI can impact supply and affordability within the housing market. The study noted the many positive changes that AI can bring to this sector, from the planning stages of real estate projects to the property tax assessments on the finished developments. But for every positive, there seemed to be a negative.
For example, in the mortgage origination process the use of AI could potentially improve the accuracy of credit ratings, particularly in seeking to redefine credit worthiness by using alternative income sources beyond the traditional credit scoring components beyond the traditional credit history sources. But the Bipartisan Policy Center also warned the technology could create more problems than solutions.
“Incorporating new data sources when evaluating creditworthiness could pose privacy and discrimination risks,” the study said. “Even if a model relies on data that does not contain characteristics protected by fair-lending laws—such as gender, religion, or race—the algorithm might instead find and use data that correlate closely with those characteristics. Alternatively, the model could be based on historically biased data. Both situations could result in unintentionally discriminatory lending decisions.”
Last week, the U.S. Department of Housing and Urban Development (HUD) released two guidance documents addressing the application of the Fair Housing Act to areas where the use of artificial intelligence poses particular concern. It’s an important start, but why hasn’t HUD taken greater and wider steps forward in considering the role of AI in housing?
HUD’s slow movement can be traced to the Biden administration’s overall lethargy on the topic. It wasn’t until last October when the administration issued an Executive Order to establish new standards for AI safety and security. While that sounds noble, it created a backlash from many tech companies because part of the Executive Order cited the Defense Production Act in claiming some AI projects posed potential national security threats – and as a result, tech companies will be required to provide information to the Department of Commerce regarding their advanced AI projects that require immense computing power.
“This is a clear violation of executive authority,” said Carl Szabo, vice president and general counsel for tech trade association NetChoice, in an interview with Politico. “I expect a swift rebuke from the courts once this gets challenged.”
Clearly, AI standards and regulations are not going to happen quickly and easily if left for Washington to formulate. In the housing sector, perhaps it would make sense if the trade organization representing the various stakeholders collaborate on their own standards and regulations that would be binding to all the companies operating therein. Rather than follow the Buffett approach of folksy ignorance or the Biden approach of my-way-or-the-highway on defining standards, the private sector should take the lead in clearly identifying where AI can go awry and creating self-governing rules to ensure the technology offers solutions and not problems.
