This article examines the importance of oversight and verification in AI systems, particularly regarding tools like ChatGPT and their impact on information reliability. The piece emphasizes that as AI becomes more influential, mechanisms to monitor and validate these systems are critical to maintaining public trust.
Trust is important. Verification is essential.
More than 2,000 years ago, the Roman poet Juvenal posed a timeless question: Quis custodiet ipsos custodes? — “Who will guard the guards themselves?” or “Who watches the watchers?”
The Romans could never have imagined artificial intelligence, facial recognition, or nationwide networks of automated license-plate readers. Yet their question may be even more relevant today than it was three millennia ago.
Recent debate over Flock Safety cameras illustrates why.
The cameras have become an invaluable law enforcement tool. A stolen vehicle that crosses several jurisdictions can now be identified in minutes rather than days. An abducted child can sometimes be located before the child reaches another state.
Investigators can reconstruct a suspect's movements at speeds unimaginable only a generation ago. These advances are remarkable and save lives and solve crimes.
Few Americans would willingly return to the days when solving a crime depended solely on eyewitness accounts or grainy surveillance footage from convenience stores.
Yet reports have surfaced that individual officers allegedly used these systems to monitor former spouses, girlfriends, or acquaintances for purely personal reasons. If true, those actions are not an indictment of the technology or of law enforcement as a whole. They are a reminder of a far older truth: every concentration of power invites the possibility of abuse.
Throughout history, every transformative invention has brought both promise and peril.
Fire allowed our ancestors to cook food, survive harsh winters, and build civilization. It also destroyed cities.
Water sustains every living creature on Earth. Floods and tsunamis destroy entire communities. Even drinking too much water can be fatal.
Electricity powers hospitals, schools, factories, and our homes. It can also be deadly.
Nuclear physics brought us clean electrical power capable of supplying entire cities. It also produced weapons capable of destroying civilization.
Full story reconstructed from Americanthinker.com. Formatting and media may differ from the original.
Artificial intelligence can help physicians diagnose disease, assist students with learning, help travelers navigate foreign countries, and help writers organize their thoughts. It can also generate convincing misinformation, fabricate legal citations, or automate fraud on an unprecedented scale.
The technology itself is rarely the issue.
The people using it—and the safeguards governing them — are.
Unfortunately, discussions about emerging technologies often become polarized. One side insists the technology should be embraced without hesitation, while the other argues it should be feared or even prohibited.
Whether we like it or not, modern life is already built on surveillance and digital tracking. Every airline passenger presents identification and is photographed. Smartphones constantly reveal our location. Credit cards, toll roads, and countless online transactions leave permanent electronic footprints. The idea that Americans can somehow “go off the grid” is largely a nostalgic fantasy.
The real question is not whether technology exists.
It is who controls it, how it is used, and what safeguards prevent misuse.
The risks posed by powerful technologies generally fall into three categories: malicious misuse, human error, and technological flaws.
A rogue police officer abusing a surveillance database constitutes malicious misuse. An investigator entering the wrong license plate or drawing the wrong conclusion is human error. Even when everyone acts in good faith, facial recognition software, artificial intelligence, and other complex systems remain imperfect.
Perhaps an even greater danger is assuming that sophisticated technology is somehow infallible.
Artificial intelligence occasionally invents facts. Attorneys have embarrassed themselves by submitting legal briefs that cite entirely fictitious court cases generated by AI. The technology was impressive, but it simply wasn’t correct.
Facial recognition has improved enormously over the past decade, but no algorithm is perfect. A poor-quality image, unusual lighting, or an unfortunate similarity between two individuals could generate a false investigative lead.
Modern medicine offers a useful analogy. No competent physician makes a diagnosis based solely on a single laboratory value or imaging study. We correlate laboratory results with the patient’s history, physical examination, and additional testing. A screening test points us toward a diagnosis; it does not establish one beyond question.
Technology should inform human judgment, not replace it.
The same principle should apply to surveillance technologies. A facial recognition match or license plate detection should trigger an investigation, not conclude one.
Nor should every concern be treated as an attack on law enforcement.
We do not conclude that all physicians are incompetent because one doctor commits malpractice. Nor should isolated misconduct by individual officers serve as an indictment of the thousands of honorable men and women who wear a badge every day and use these technologies responsibly to protect their communities.
In medicine, every access to an electronic medical record is logged. Hospitals routinely audit those records. Years ago, hospital employees made headlines for improperly accessing the medical records of celebrities out of curiosity. Those audit trails made the misconduct discoverable, and consequences followed.
If surveillance systems maintain similar audit logs — as Flock reportedly does — the obvious question is: Who is reviewing them?
An audit trail is valuable only if someone conducts the audit.
Imagine a system in which every database search is permanently recorded, random audits are routine, unauthorized access results in suspension or termination, and deliberate misuse carries criminal penalties. Public confidence would increase dramatically without diminishing the technology’s legitimate usefulness.
The Framers of the Constitution understood this principle long before computers existed. They didn’t design a government on the assumption that those in power would always act honorably. They designed it around checks and balances because they understood human nature.
James Madison famously observed that if men were angels, no government would be needed.
Perhaps the greatest mistake we can make is demanding perfection. No technology created by human beings will ever be flawless. Airplanes occasionally crash despite extraordinary safety records. Nuclear reactors fail despite layers of engineering safeguards. Self-driving vehicles make mistakes. Human beings certainly do.
The answer is to build systems that anticipate imperfection.
Good engineering assumes that components will fail.
Good medicine assumes that diagnoses can be wrong.
Good government should assume that power, if left unchecked, will eventually be abused.
Good aviation assumes that pilots can make mistakes.
The challenge before us is therefore not whether to embrace new technologies. That decision has already been made. Flock cameras, artificial intelligence, facial recognition, drones, and innovations we have not yet imagined will become increasingly woven into everyday life.
The enduring question is the one Juvenal posed twenty centuries ago.
The answer cannot be simply “trust them.”
It must be Ronald Reagan’s famous advice, “Trust but verify.”
Powerful technologies deserve equally robust oversight. That oversight protects not only individual liberty but also the integrity of the institutions that use these remarkable tools. When the public knows there are meaningful safeguards against abuse, confidence grows rather than eroding.
Progress without accountability breeds distrust. Accountability without progress breeds stagnation. Liberty requires both innovation and oversight.
Brian C. Joondeph, MD, is a Colorado-based ophthalmologist who writes frequently about medicine, science, and public policy.
Now more than ever, the ability to speak our minds is crucial to the republic we cherish. If what you see on American Thinker resonates with you, please consider supporting our work with a donation of as much or as little as you can give. Every dollar contributed helps us pay our staff and keep our ideas heard and our voices strong. Thank you.
