A county has ended a license-plate surveillance contract amid privacy concerns, while a national poll finds most Americans think AI is moving too fast. Both stories put pressure on organizations to show how they control data use, not merely describe what their technology can do.
Henrico County terminates Flock surveillance contract amid privacy concerns
Henrico County has ended its contract with Flock, which operated 122 license plate readers, amid concerns about data protection and potential misuse. The decision follows reported incidents of unauthorized use by law enforcement and scrutiny of similar systems in other jurisdictions. For public-sector buyers, the issue is not just whether a reader captures useful information; it is who can search that information, for what purpose, and under whose supervision.
Ending a contract is a consequential governance response because it removes a deployed system rather than asking the vendor or agency to improve its practices while operations continue. The reported unauthorized use also points to a distinction procurement teams should make explicit: a tool can function as designed while its access rules fail to prevent improper use. Buyers evaluating surveillance products need to assess the operating controls alongside the hardware and software.
- Procurement teams should test whether permitted uses, user permissions, and independent oversight are defined before a surveillance system goes live.
- Data leads should require access logs that make improper searches detectable and give reviewers enough context to investigate them.
- Privacy and compliance teams should specify retention, sharing, and deletion obligations in contracts rather than leave them to informal operating practice.
- Vendors serving multiple jurisdictions should expect concerns about misuse in one deployment to prompt questions about controls in others.
AP-NORC poll: Most Americans say AI is advancing too quickly
An AP-NORC poll found that 64% of Americans believe AI is advancing too quickly. Respondents raised concerns about privacy, job displacement, and wider societal effects, and the poll found limited trust in political leaders to regulate AI effectively. The result is a measure of public sentiment, not a finding that any particular AI product is unsafe or that regulation has failed.
For teams deploying AI, that distinction matters: the poll does not prescribe a technical fix, but it does identify questions users and customers are likely to ask. They may want to know what data a system uses, whether personal information is exposed, how decisions can be challenged, and who is accountable when it causes harm. Low confidence in political leadership makes clear, verifiable explanations of organizational safeguards more important, even though those safeguards cannot substitute for public policy.
- AI teams should document data sources, access restrictions, and review procedures so privacy assurances can be checked rather than taken on trust.
- Product leaders should address potential effects on workers and affected users when explaining a deployment, not treat privacy as the only concern.
- Compliance teams should distinguish what their controls demonstrably prevent from broader claims that an AI system is responsible or safe.
- Organizations should give users a clear route to question outcomes or report problems, because visible accountability matters when confidence in oversight is low.