A Safer University Campus Without a Culture of Blanket Surveillance

Why blanket surveillance is the wrong default for a campus

A campus is a community built on open debate, experimentation, and the right to question authority. When cameras and analytics multiply faster than evidence of their benefit, that foundation erodes. At the University of Michigan, the installation of more than 1,200 new security cameras, some with audio capability and placed in public gathering spaces, prompted a faculty senate resolution: the proliferation of surveillance technology, members argued, compromises the freedom of faculty, students, and staff, and they called for a funded study of whether cameras actually deter and solve violent crime before more units are added.

The University of Melbourne case shows where this drift leads. Wi-Fi location data collected for network operations was later used to identify students who attended a sit-in protest and to support misconduct proceedings. Victoria's privacy regulator found this breached core privacy principles: students had not been told how their location data could be used, and the surveillance purpose was not one they could reasonably have expected from simply connecting to campus Wi-Fi. The lesson for any institution: before adding a sensor, demonstrate that a specific risk in a specific zone calls for observation rather than for lighting, guards, or access control.

Existing research on whether cameras deter crime is thin and dated. Coverage, in other words, grows faster than proof of value. Treat an absence of demonstrated need as a reason not to install.

Start from a written purpose, not from camera count

The most reliable safeguard is a policy that classifies every camera by its function. A well-designed institutional policy typically sorts cameras into three purposes: property protection (capturing theft or damage in labs and parking lots), personal safety (capturing an assault in a walkway or lounge), and extended responsibility (live monitoring by nearby staff in food service areas or testing centers). Each camera should map to one of these stated reasons and to a location that justifies it.

Equally important is what cameras may not do. Strong policies state that surveillance will not be installed to conduct personnel investigations — for attendance, work quality, or academic conduct — though recordings captured in routine operation may still be used if there is reasonable cause. Some policies go further and rule that material gathered in violation of the policy cannot be used in a disciplinary proceeding against a student or employee. This is the boundary that keeps a safety system from quietly becoming a performance- and behavior-monitoring system.

Write the purpose and the prohibited uses before the procurement. A camera bought with no articulated purpose is a problem looking for an excuse to exist.

Set rules for placement, retention, and access before installation

Placement rules protect the spaces where a person retains a reasonable expectation of privacy. Mature campus policies prohibit cameras in student living spaces in residence halls, bathrooms, and locker rooms; they limit views into residential windows; and they ban dummy or placebo cameras, requiring that every unit be visible and actually operational. They also generally prohibit audio recording. Where private residences or housing sit within camera view, the framing should be adjusted to avoid intruding on them.

Retention and access are where most operational risk hides. Rather than storing footage indefinitely because disk space is cheap, set a specific retention period — roughly 30 days is a common, defensible default for higher education — and delete automatically. This protects the institution when a public-records request arrives months later: if a documented policy limits retention, you cannot be expected to produce footage that no longer exists. Access to live and archived video should be limited to authorized, trained personnel; default operation should be passive recording rather than continuous live monitoring, with live monitoring reserved for high-risk zones, restricted areas, alarms, special events, and authorized investigations.

Keep an access log of who viewed or exported footage, restrict copying and retransmission to the police unit, and require VMS users to complete training in the legal and ethical boundaries of camera use. Each of these steps converts an abstract privacy promise into an auditable control.

Build standing oversight and transparency

Surveillance becomes defensible when it is governed by a procedure rather than by one department. The model used by leading institutions is a standing oversight committee that includes the chief of police or security, the CIO, student affairs, human resources, and facilities — and, in many cases, elected faculty and student representatives. Such a committee approves camera requests, reviews complaints and appeals, and periodically re-examines the policy. Community-facing institutions publish a map or list of where cameras sit, what they record, how long footage is kept, and how anyone can ask a question or file a complaint.

Transparency serves two ends. It reassures the community that safety does not mean being watched like suspects, and it disciplines the security operation itself. When oversight, review, and published summaries are in place, mission creep — the slow migration of a camera from its stated safety purpose toward monitoring individuals or groups — becomes far harder. When cameras were installed in public gathering areas at Michigan, the ACLU and faculty governance objected precisely because students were not included in the required oversight committee and the policy had been weakened without public review. Standing oversight is the mechanism that would have caught that drift early.

Publish at least annually: the count and rough locations of cameras, stated purposes, retention periods, who may access footage, and how to exercise rights of review or complaint.

• Confirm whether facial recognition or biometric collection is legal and consent-compliant in your jurisdiction
• Deploy advanced analytics only in public areas where there is no reasonable expectation of privacy
• Ensure AI tools store anonymized data only and for limited timeframes
• Require human review for any automated alert rather than autonomous action

Treat each advanced technology as a separate, documented decision

Facial recognition, automated license plate readers, AI weapon detection, and biometric access control are not natural extensions of a camera network — each is a distinct decision with its own legal and ethical weight. Facial recognition collects biometric data and is banned or restricted in several cities and states; biometric authentication is best reserved for narrow, high-security applications such as labs with hazardous materials, and only with informed consent. Public universities also face open-records exposure: recording public streets and sidewalks can obligate them to release footage of incidents that never touched their property, so limiting camera fields to building entrances reduces that burden.

There are proven, proportionate alternatives. Some institutions decline facial recognition entirely and instead use analytics on anonymized data with short retention, or similarity search that analyzes height, clothing, or gait without storing biometric records. License plate readers can be deployed so the university owns the data, retains it only as long as state law requires, and grants no outside access. AI weapon detection tools work with human analysts around the clock on anonymized feeds. Before adopting any of these, define the specific use case, the jurisdiction's legal constraints, the data flows, and the retention schedule — and be ready to say no when the benefit is speculative.

Define what happens when police or outside bodies ask for footage

The hardest governance question is disclosure. Cameras exist partly to help solve crimes, but easy access to footage of assemblies and marches can be turned against people for their lawful expression. Civil-liberties groups have repeatedly warned that sharing campus camera feeds with city real-time crime centers increases the risk that footage of speech, or of students' immigration and political activity, will be used for targeted action.

A responsible policy keeps disclosure controlled and recorded. Footage is treated as confidential and used only for official university and law-enforcement purposes; any release to an outside agency is reviewed by a designated official with legal counsel rather than made available through an open feed. The policy should state plainly that surveillance is not used to target individuals or groups for their beliefs or affiliations, and that participation in peaceful assembly is not itself grounds for reviewing footage of a specific person. This is the point where a safety system either earns trust or loses it.