Blockchain, AI will reshape real estate
Innovations in computing and information technologies are transforming businesses at an unprecedented pace. The real estate sector will not be spared, and that’s good news.
Advances in artificial intelligence, analytics and blockchain will improve the real estate industry by achieving greater efficiencies and identifying and mitigating risks.
While real estate will remain true to its traditional brick-and-mortar roots, the technological innovations will transform the way the sector operates.
Consider blockchain, which offers great promise for transparency, data accuracy, and data aggregation. Essentially, a blockchain is a “distributed database that maintains a growing list of data items and that is hardened against manipulation and counterfeiting,” explained Jan Veuger in the journal Facilities in 2017.
Once a block is added, it acts as a “building passport.”
It allows one to aggregate all relevant information about a building in one place from ownership records to structural details down to the latest renovations.
Such information can be made readily available to those interested.
Another example of technological innovation is valuation models.
For centuries, the value of a real estate asset was determined by expert opinion that involved examining earlier comparable sales, applying some judgments and coming up with an estimate.
The process was error-and bias-prone.
The practice was transformed in the past few decades when statistical models effectively replaced the expert estimates.
The use of computer algorithms in valuation became the norm. Advances in data storage and computing powers meant that statistical models improved in predictive accuracy.
Recent advances in analytics imply that even more sophisticated computing algorithms will soon be the norm in the valuation space.
Whereas statistical models of the regression type are common today, the future will see a wider application of machine learning algorithms including Artificial Neural Networks (ANN) and Support Vector Machines (SVM).
Computational advances are likely to transform the mass appraisal market that involves determining the values of groups of properties at a given time.
Big users of mass valuation models are public sector entities responsible for property taxation.
The property tax is calculated for all properties at a given date, which requires estimating the value of each property in the tax roll. Not getting the valuation right could lead to expensive litigation costing millions.
The other big users of mass valuation are mortgage lenders who would like to ascertain the value of a property before extending the loan. Again, getting the value wrong could expose the lender to greater risk.
Recent research shows that machine learning algorithms offer improved performance for predictive analytics.
The gains over the traditional regression type models are even higher when data depict non-linearities.
Already, numerous startups have emerged in the U.S. and Canada trying to get a piece of the automated valuation business in real estate.
The startups have met with varying degrees of success in securing venture capital.
The real success though is their ability to deliver an accurate valuation of real properties. For this, they must rely on not just the best algorithms but also the best data.
Garbage in, garbage out applies to the AI world as well. An AI or machine learning model learns from the data we feed to train the algorithm. Poor quality data means poor training and inferior forecasts. Thus, the future success of predictive modelling is incumbent on improving techniques to weed out outliers and erroneous data.
AI, therefore, needs blockchain to access high-quality property data.
The advances in AI models are focused on replicating the workings of a human brain. It is rather odd that computer-based models were introduced earlier to replace human decisionmaking with algorithm-based tools.
The future is far from certain. If computers can think like humans, will they make the same cognitive mistakes that humans make? Or being artificially intelligent, will computers be able to mimic human decision-making without being swayed by emotions?
Real estate transactions, especially housing, may never be devoid of emotions.
If AI implies intelligence without emotions, the valuation models may depict a greater variance between estimates generated by humans and computers.
It’s free will and the readiness to be swayed by emotions that separates humans from robots. AI-driven automated valuation models are fast improving in statistical intelligence. Emotional intelligence, though, is hard to machine learn.