this has been true since photoshop existed? you don't need an algorithm to make a fake picture or to fake a military attack, etc. humans generally have better grasp of appropriate details
The biggest problem with this line of response is not only that it utterly ignores scale, cost and speed, as has already been addressed in responses, but consequence.
Yes, in the 8th century BCE one could commission bards to gin up epic sagas of conquest and atrocity to stir the public spirit. In the 15th century, one could hire one of the most brilliant sculptors and painters of all time to create artpieces legitimising a power-hungry thieving dynasty which had seized control of banking, commerce, and the Church. In the 16th century, an earlier horde could by mobilised by an Orange tyrant through pamphlets cheaply printed and distributed. In the 19th century, Wreck-a-Feller could intercept (and rewrite) telegraph transmits, newspapers could sensationalise photographs, and preachers could roll holy until districts burned out.
And so into the 20th and early 20st centuries.
What was not possible until quite recently was for high-fidelity, all-but indistinguishable, continuous imagery, audio, and video to be produced in realtime and distributed globally.
The time between fraud and mass impact is now measure in seconds, and can reach globally. This is well within response intervals not only of populations and institutions, but of individual weapons of mass destruction. The risks of massive consequence are huge.
And that is what is new and novel.
The interaction with other technologies raises yet more concerns. With Slaughterbots a highly-plausible if not actual reality, "leaked" video showing a plausible Slaughterbots attack might itself be a weapon of war. (https://yewtu.be/watch?v=vR91F3tp6eQ)
That is, one of the risks of some presumed capability or knowledge is that even the belief of its existence or validity becomes a tool to be used. The late senator from Wisconsin claimed to have, though never presented, evidence of disloyalty, a common practice in witch-hunts. With increasing scope and scale of data collection and storage, such claims become of themselves ever more plausible, with numerous effects. Concepts such as the Panopticon and "chilling effects" operate not based on actual or observed risks, but on presumed ones. Adam Smith in Wealth of Nations uses the word invisible twice. The unhanded one references the decline in moral amongst soldiers faced with sudden and invisible death at any time, and the costs of perpetual vigilance.
So, no, really, this is not the same. Scale matters. Effects matter.
It's a matter of quantity as well as quality. It gets easier to produce convincing forgeries all the time, but it doesn't get easier to debunk them. This allows for public discourse to be destroyed by flooding it with fakes.
