For an obvious example, governmental financial data now stored in various databases will hold keen interest for municipal bond investors. Such information is collected now by the Municipal Securities Rulemaking Board in its electronic database. Other organizations own complementary pieces of the puzzle, which could have commercial value when combined by an analytical engine to provide insights into potential bond ratings changes, default risks and subtle improvements in financial measures. Those could change the market valuation of billions of dollars of tax-exempt securities. And that is just one use case.
Other state and metropolitan public information sources are likely to offer similar treasure troves of raw data that can now be compiled into analytical warehouses by scraping information from public websites. There are likely to be dozens of other types of governmental data files routinely maintained for internal use, but not presently posted on the Internet, which could still be searchable under Freedom of Information (FOI) requests.
Just how the various states and localities manage a potential future wave of commercial requests for public information in electronic form is still an emerging issue. Today, some states and localities authorize only in-person, on-site viewing of public records, which will hobble the speed at which these new technologies take hold in the public square.
But it seems inevitable that electronic data files will increasingly become part of tomorrow’s public-sector interfaces, drawing the interest of information vendors in the AI-driven database industry. No doubt there’s a lawsuit eventually coming somewhere on this issue of commercial access to public data.
The first challenge for public administrators in this new race for information access will be to identify which public records files that they now manage might contain information of value to commercial interests. Public health agencies, public schools, state education departments, law enforcement agencies, toll road authorities, property registrars and building inspection departments are likely to house records that could conceivably contain useful nuggets when systematically compiled by an AI-powered computerized learning program to glean actionable insights into trends and markets.
Needed: New Ground Rules
State legislatures and their support staffs will need to revisit their own Freedom of Information laws in light of the AI revolution. Artificial intelligence has already shown up on legislators’ radar in the context of election laws, but that’s tangential to this sphere of issues. New FOI debates and resulting laws will focus less on paper-based documents and more on extracting and assembling digital bytes of information. Just as gold is now mined commercially by sifting through huge heaps of rough ore with chemical leaching, public information files will be reprocessed by AI ingestion and ever-evolving algorithms.
At a minimum, each state will need to lay out the ground rules for accessing electronic data files housed by public agencies. That should include what are known as “terms of use” and possibly even standardized licensing agreements that would place restrictions on how those data files are compiled, priced and used, in some cases including a royalty obligation on the commercial users of such data.
At the national level, the governmental professionals working in this field, such as the experts at the National Conference of State Legislatures and the National Association of State Auditors, Controllers and Treasurers, would do everybody a great favor by quickly assembling some model legislation to help the states and localities navigate through this inflection point in the history of public information laws.
Industry lobbyists will surely argue that taxpayers have already paid for the data and that it therefore should be made available to private parties for nothing more than the costs of data retrieval and transfers. But that ignores the actual value of the data and importantly, which taxpayers and residents have previously paid for its collection, retention and transfer: Most requests will likely come from organizations located outside the jurisdiction, and even if the data seekers are local, the financial benefits of this information should be shared with the data collectors at fair value and not just the jurisdiction’s marginal cost of uploading it. Otherwise, our society runs the risk of privatizing public data for the benefit of a new legion of American AI oligarchs.
Identifying Hidden Assets
The various professional associations that serve state and local governments also will have a ripe opportunity in the coming year to raise awareness among their membership that the AI revolution could be coming to their desks soon and help identify the most likely data fields that private companies would seek to harvest for commercial purposes. Just knowing where there may be hidden assets is the first step to protecting them; then the challenge will be to estimate their potential value in financial terms.
It’s too early to start booking these as assets on the balance sheet, but maybe a future financial statement footnote should disclose the nature and dimensions of such potential hidden “intangible assets.” In most cases, the commercial value of the data will not become a dominant line item in a state or municipal budget, but it could someday have fiscal value greater than the traditional revenues derived from business licenses, for example. Taxpayer groups will grouse if public officials let economic value slip between their fingers.
If there is to be a wave of public information requests from the AI industry, the initial beachheads will be the states and the largest municipalities. Just as “Slick Willie” Sutton robbed banks because “that’s where the money is,” so too will the database builders start where the largest, most potentially valuable files reside. That will give plenty of time for smaller jurisdictions to prepare and learn from the experience of others.
In many public organizations, the first responders on staff may be in the information technology department, as they would be well aware of the various data files kept by each entity. It is most likely that the IT staff will become involved at some point when FOI requests arrive at the desks of operating agencies and legal departments.
In some cases, it is clearly conceivable that data files requested by AI-assisted data miners will contain personal information that must be scrubbed from the files that would be transmitted to a private party. State laws will likely govern that process, if state laws are silent or insufficient, local governments would be wise to establish their own ground rules for identity protection. Because of the intellectual property dimension at this time in history, in-house attorneys should be made aware of these emerging issues and give thought to the best ways to protect individuals’ information as well as their agency’s potential property rights and data business interests.
The Role of the CFO
At a minimum, and even if they are not assigned to take point in this research, governmental finance officers need to acquaint themselves with the potential data sources that could someday have commercial value to assess whether there could eventually be potential revenue streams. In this situation, nobody needs to act alone, because the private sector will not stop at a single agency if companies sense a commercial opportunity. CFOs should huddle with regional and national peers who have also begun to think this through. Inevitably this issue will arise in regional and statewide professional associations and chat groups in the coming year, and it cannot hurt for the finance officers to make the jurisdiction’s top officials and senior staff aware of these emerging issues.
Finance officers can also help their IT and operating department counterparts deal with the inevitable claims that AI systems will magically enable teachers, police, building inspectors and other public workers to reduce their paperwork time in miraculous ways that save taxpayer money. Sure, someday firefighters may have AI-smart vision enhancers installed in their helmets to guide them inside burning buildings by using blueprint and inspection imagery, but not in 2024 or 2025. Other shiny new objects will be coming soon. But right now that’s mostly a lot of anticipatory hype, and the CFO’s role in many cases will be to help manage expectations and call out boondoggles that over-enthused early-adopter department heads may naively want to buy into. As the investing guru Daymond John put it, “Pioneers get slaughtered, and settlers prosper.”
The AI industry is not yet performing productivity miracles, although some will come eventually. Clearly, the data miners are not yet fully equipped to begin an assault on public-sector data files, and in many states the limitations of FOI laws may present an initial hurdle that provides some prep time to get ahead of the next wave of industrial innovation.
But the history of technological change is that it will be impossible to put the genie back in the bottle if machine learning systems can be assembled to produce a marketable end product for commercial use of now-underpriced public data. Today that’s all blue sky, but there will be millions of dollars invested in coming years to try to find novel ways to perform financial alchemy from public records. Only time will tell which of them actually pan out, whether they produce gold — or silver or copper — and who benefits.
Governing’s opinion columns reflect the views of their authors and not necessarily those of Governing’s editors or management. Nothing herein should be construed as investment advice.
Related Articles
