Paid actors turned political outrage into a repeatable video business

The presenters looked like ordinary people delivering political commentary from home. Semafor found that several were paid actors reading repetitive, apparently AI-assisted scripts, often containing misleading or false claims about Democratic politicians.

One interconnected network attracted more than 45 million YouTube views. The investigation traced a production model associated with growth entrepreneur Caleb Chan: create channels, hire inexpensive spokespeople, identify attention-grabbing subjects and turn audience traffic into advertising revenue. Human faces supplied the appearance of independent judgment while scripts and distribution could be manufactured at scale.

That makes the phenomenon different from an entirely synthetic avatar channel. The real person on camera does not necessarily originate the claims or represent an independent editorial operation.

In its follow-up, updated September 7, Semafor reported that YouTube terminated 20 channels, including all 13 named in the investigation and seven related accounts, for violating spam rules. Chan confirmed details of the reporting while disputing its interpretation: he described an ideologically motivated media studio and called the coverage unfair.

The enforcement removed particular accounts. It did not establish that the underlying economics of cheap, personalized-looking political videos had disappeared.