News
WSAZ-TV reports that InvestigateTV+ examined how fraudsters build convincing AI voice clones from ordinary people's speech in seconds, often unnoticed.
Reported by
WSAZ-TVAccording to WSAZ-TV, the investigative unit InvestigateTV has published a report examining how fraudsters use AI technology to create voice clones in a matter of seconds. The report, part of the InvestigateTV+ series, looks at how online criminals are profiting from the technique and lays out steps families can take to protect themselves.
The report states plainly that a person's voice may no longer be something only they can produce. Scammers have found ways to capture the sound, tone, and inflection of everyday people's voices and turn them into AI deepfakes, often without the person ever knowing their voice was taken.
InvestigateTV+ is a journalism unit, not a law enforcement body. What it did here is put a name and a shape to a threat that many people have not yet encountered: cloned voices built from ordinary speech, used to defraud. That naming matters on its own. A reader cannot guard against a technique they have never heard described.
The report frames the problem in specific terms: seconds, not hours, to build a convincing clone; ordinary Americans, not public figures, as the source material; and a victim pool that may not know a clone of their voice exists at all. That last point is the one worth sitting with. A stolen wallet is missed within a day. A stolen voice might never be missed, because the person it belongs to has no way of knowing it was taken.
The same InvestigateTV+ segment, in the same publication cycle, covers a separate story: more than 2,500 people in Tennessee arrested for DUI between 2017 and 2024 who were later shown by bloodwork to have had no alcohol or drugs in their systems, part of a pattern spanning 22 states. That story is not about voice cloning. It shares a publisher and a segment with this one, and it shares a lesson: a test or a technology can carry the appearance of certainty long before it has earned it.
Publication is the first step in any consumer protection effort. Regulators, banks and telecom carriers cannot build defenses against a fraud pattern that has not been documented and named. InvestigateTV's report gives family members, particularly those who might be targeted through a cloned relative's voice in a distress call, language for what to ask about and watch for. The report does not name a law enforcement action or a company held accountable. What it establishes is that the technique is active now, it targets ordinary voices, and it moves fast enough that awareness has to come before the call, not during it.
