If institutions demand that citizens accept extensive monitoring while resisting any examination of their own surveillance practices, they risk creating the impression that privacy is a privilege reserved for those who hold power.
Conversely, if private individuals dismiss all concerns about tracking public officials, they risk ignoring legitimate safety issues.
A mature public debate should reject both extremes. It should recognize the importance of public accountability while acknowledging that the collection and dissemination of location data can create real risks.
Sistilli's project does not settle the question of which surveillance practices should be permitted. Its value lies in forcing people to articulate the principles they believe should govern them.
When the roles of observer and observed are reversed, assumptions that once seemed obvious become open to examination. That is one of the experiment's most important contributions.
10. Transparency and Accountability in the Age of AI
The spread of automated surveillance raises a broader question about how public institutions and private companies should explain the technologies they use.
When surveillance systems operate without clear public information, people may struggle to understand what is being collected, how long records are retained, who can access them, and whether the information is shared with other organizations.
These uncertainties can undermine trust even when a system has legitimate uses.
Transparency does not necessarily mean publishing every operational detail. Some information may need protection to prevent criminals from evading detection or to safeguard sensitive operations. However, the basic rules governing data collection and use should be understandable to the public.
A meaningful transparency policy might explain the purposes for which a system is deployed, the categories of information it collects, the retention periods that apply, and the circumstances under which records can be searched or shared.
It could also describe the procedures used to address errors, review complaints, and assess whether the system remains necessary.
Accountability requires more than making information available. It also requires mechanisms for evaluating conduct and responding when rules are violated.
For example, an organization operating a vehicle monitoring system should be able to establish who accessed the database and for what purpose. Access logs can help identify inappropriate searches, while independent audits may reveal whether the system is being used in ways that exceed its stated objectives.
Clear responsibilities are also essential. If a company provides the technology but a public agency operates it, both parties may have roles in protecting data and ensuring that the system functions appropriately. Agreements should clarify those responsibilities rather than leaving important questions unresolved.
The same considerations apply, in a different form, to privately operated surveillance projects.
An individual who creates a monitoring system should consider whether the collection is lawful, whether it could expose people to harm, and whether the information needs to be stored or shared. Technical ability should be accompanied by an understanding of the consequences.
AI can make surveillance more powerful because it reduces the effort required to identify patterns and process large quantities of information. But the existence of an automated capability does not determine whether using it is justified.
That decision requires ethical, legal, and social judgment.
The challenge is particularly difficult because AI systems can appear neutral. A camera records an image, software identifies a plate, and a database stores a result. Each step may seem mechanical, giving the impression that the overall process is objective.
In reality, people make choices at every stage. They decide where cameras are placed, which vehicles should be identified, what information should be retained, how results should be interpreted, and who should have access to the database.
These choices influence which activities become visible and how the resulting information can be used.
Accountability should therefore extend beyond the software itself. It should include the institutions and individuals who determine the system's purpose, configure its operation, and act on its outputs.
Sistilli's experiment makes this point in a simple but powerful way. The same technical process can support different goals depending on who controls it. Technology does not remove the need to evaluate those goals.
11. Could Automated Surveillance Improve Public Safety?
Despite the concerns surrounding privacy, automated license plate recognition has potential benefits that should not be ignored.
When appropriately designed and lawfully deployed, such systems may help investigators locate vehicles associated with serious incidents, identify possible connections between separate events, or reduce the time needed to review large amounts of footage.
Automation can be particularly useful when the volume of available information exceeds what people can reasonably examine manually. A searchable database may help investigators narrow a large set of observations to a smaller group that warrants further review.
In this respect, the technology can improve efficiency. It may allow personnel to spend less time on repetitive identification tasks and more time evaluating evidence, conducting interviews, or pursuing other investigative work.
However, these benefits depend on the circumstances. A system that produces inaccurate results, collects excessive data, or lacks adequate oversight may create costs that outweigh its advantages.
The existence of a useful capability does not automatically justify its use everywhere.
A central policy question is whether the surveillance system is proportionate to the problem it is intended to address. A targeted system used for a defined investigative purpose may be easier to justify than indiscriminate collection of information about every vehicle passing through a large area.
The distinction matters because broad collection can affect people who have no connection to the activity that prompted the surveillance.
A policy that emphasizes necessity would ask whether the same objective could be achieved through a less intrusive method. It would also consider whether the information collected is relevant to the stated purpose and whether retaining it for an extended period is necessary.
The design of the system can help support these goals. Access restrictions, retention limits, audit trails, and clear procedures for handling search requests can reduce opportunities for misuse.
Human oversight is equally important. Automated results should support decision-making rather than replace careful evaluation of evidence. Users need to understand that a possible plate match is not necessarily proof of a person's involvement in an incident.
Public safety and privacy should not be treated as competing objectives that require society to choose one and abandon the other. In many cases, a well-designed system can seek to achieve legitimate safety goals while limiting unnecessary collection and exposure of personal information.
Nevertheless, a balance cannot be established through technical design alone. It also requires public debate about which purposes are legitimate, what safeguards are sufficient, and who has the authority to make those decisions.
Sistilli's project contributes to this debate by making the underlying capabilities visible. It encourages the public to examine both the potential usefulness of automated monitoring and the risks associated with its expansion.
12. The Risk of Function Creep
One of the less visible risks associated with surveillance systems is function creep: the gradual expansion of a technology beyond the purpose for which it was originally introduced.
A system may initially be deployed to address a narrow problem. Over time, however, users may identify additional applications for the data, new organizations may request access, or the system's capabilities may expand through software updates and integration with other databases.
Each change may appear reasonable when considered separately. The cumulative result, however, can be a monitoring system that collects or uses far more information than originally anticipated.
For example, a vehicle recognition system might begin as a tool for investigating specific incidents. If its database gradually becomes available for a wider range of searches, information collected for a limited purpose may be used in circumstances that were not part of the original justification.
The issue is not that every expansion is necessarily harmful. New uses may provide genuine benefits. Rather, the concern is that each expansion should be evaluated on its merits instead of being accepted simply because the technology is already in place.
Sistilli's experiment highlights the importance of this question because it separates the basic technical function from the purpose assigned to it.
The same camera and recognition software could be used to record police vehicles, monitor access to a private facility, or support a narrowly defined investigation. The technology itself does not determine which purpose is appropriate.
As the range of possible applications expands, the rules governing use become increasingly important.
Organizations should establish clear boundaries before deploying a system and review those boundaries when the system's purpose changes. They should consider whether new uses are consistent with the original justification, whether they require additional safeguards, and whether affected people should receive further information.
Data retention is another important part of the problem. Information that is kept indefinitely may become available for purposes that could not reasonably have been anticipated when it was collected.