Hunton Retail Law Resource

Artificial intelligence (“AI”) is often touted as the latest transformative technology set to revolutionize the commercial real estate industry. This may seem like grandstanding, but the hype is real. AI is already being implemented by law firms and their clients to streamline business processes. AI, however, is ultimately a tool and whether tools make our lives easier or harder often depends on if we use them safely. This blog post examines some of the ways in which AI is used in commercial real estate and potential pitfalls with broad implementation.

Data Synthesis and Analysis

AI platforms can quickly synthesize information in bulk much more quickly than a human. A retail client may have thousands of properties in its portfolio. The portfolio may consist of properties with varying locations, land uses, ownership structures and the like, and each property will likely have an associated file documenting this information. AI-powered databases can quickly group properties by type (such as leaseholds versus fee interests), summarize common or key terms and generate asset performance data. By utilizing AI, a retailer can significantly cut down on costs by summarizing information in a more efficient and uniform manner.

Legal Transactions

Generative AI involves the use of machine-learning algorithms to create content and work product. Large retailers often engage in “template law” for consistency. Drawing on a closed universe of existing documents, AI can be used to draft legal documents on a client’s preferred form. As mentioned above, if documents are processed in an AI-powered database, AI can also be used as a resource during negotiations to create a flow chart of common or precedent terms and historical best practices for a given situation. Transactions, however, are often littered with exceptions and market standards are hardly immutable. As AI-generated documents tend to be generic by nature, users will need to be particularly mindful of deal, property and market specifics.

Management and Operations

Many large companies already use AI to assist with asset management and property-level operations. AI can be used to generate marketing materials and property descriptions. Asset managers use analytical AI to create budgets and forecasts, and can use AI to streamline ESG data collection and reporting. On an operational level, a generative AI platform can process a freshly negotiated agreement and produce a critical dates timeline and model plans and specifications for improvements.

Potential Pitfalls

AI is a powerful tool, but AI-generated outputs are only as good as the inputs. The databanks that form the basis of an AI platform’s knowledge may contain information that is outdated or flat-out wrong. Outputs using such information run the risk of producing work or content with the same inaccuracies. Human users will also need to be trained to input appropriate prompts and review outputs. In a similar vein to researching a legal issue, to get the right answers users will need to know how to ask the right questions and filter out errors. AI is, as the name explicitly spells out, artificial intelligence—it does not possess conscious knowledge. To put it (overly) simply, a generative AI model is a hyper-advanced text predictor designed to generate coherent responses based on precedent inputs. That coherency makes it all the more dangerous when the outputs are wrong. Indeed, there have already been several high profile cases of lawyers citing fake cases fabricated by generative AI platforms. The ramifications could be similarly disastrous if a user relied on plausible but ultimately fake laws or zoning regulations created from an AI platform to push a project forward. These are only a few threshold issues with respect to implementing AI. Retailers will also need to consider issues with data privacy, copyright, IP and proprietary rights, and the all-around uncertainty with respect to regulation that comes with any new technology.