In January 2023, the San Francisco Standard wrote an article titled, “How a Fake Doctor Tricked Pundits by Fueling Covid Outrage on Twitter”.
The article shows a “Dr Robert Honeyman”, who naturally sponsors the government narrative.
The Standard asked the fake account about their various sob stories, to which the account owner admitted to being the employee of a ‘troll farm in southern Africa’, before vanishing:
It might retroactively seem obvious, hindsight is 20/20, but the general public seem to be unaware of just how sophisticated AI has gotten, and the ability to produce convincing forgeries impacts us far into the future.
Ground Rules
To show you what we mean, we’ve set up a some test examples.
To make the test fair, we discourage you from using either personal or online tools to check, because the majority of people — tech unsavvy — will not use analysis tools. Be honest, you probably don’t haze most images to the third degree anyway.
Further, we’ll use the same strategies as any misleading fraudster would: only picking high quality image samples, editing images, and even using composites (one or more real world images with a forgery inserted by AI).
Reality Check
Lets test your ability to tell reality from un-reality
Answers will be revealed at the bottom of this article.
Writing Test
AI can also write convincingly. Without looking it up, determine which is fake?
Shakespeare
Paragraph A:
"Neither a borrower nor a lender be; For loan oft loses both itself and friend, And borrowing dulls the edge of husbandry." - William Shakespeare, Hamlet (Act I, Scene III)
Paragraph B:
"Fair maiden, like the sun that breaks the dawn, thou dost illuminate my heart with a radiance brighter than a thousand stars." - William Shakespeare, Romeo and Juliet (Act II, Scene II)
Of Mice And Men
Paragraph C:
"He had the look of a man who had seen hard times and come out the other side, but the scars of his past still lingered in his eyes like shadows of regret." - John Steinbeck, Of Mice and Men.
Paragraph D:
"Guys like us, that work on ranches, are the loneliest guys in the world. They got no family. They don't belong no place." - John Steinbeck, Of Mice and Men.
What Do We Do About The Problem?
Unfortunately the solution here isn’t regulation. People will just use AI behind the scenes to generate whatever they want regardless. You can’t regulate other countries and there’s no way to be absolutely sure something isn’t AI generated.
It’s a bit like banning guns; criminals don’t care what the law says, they use them anyway. AI will be much the same. Everyone is already spinning off their own instances, the cat is already out of the bag.
We’re at an inflection point where AI content is hard to distinguish from real world. It is insulting to visit an AI site and prove to an AI we’re human via a captcha. These AI bots can be used to generate fake accounts and drive hysteria, whether in pandemics or other manufactured ‘disasters’.
These forgeries will only get more convincing as time goes on, we may even face a ‘crisis in truth’ as we become unable to distinguish what is and is not real online.
Did you pass the test dear reader?
Answers To Test:
Image B is from here (we don’t know if it’s real).
Image D is from here (we don’t know if it’s real).
Image E is from here (real photograph on PixaBay).
Paragraph A is real.
Paragraph D is real.
Note: The AI generation only has to induce enough uncertainty in choice to ‘win’, not to become ‘more convincing’.
Subscribe to receive more content from The Daily Beagle.
Share to let the world know!
Or leave a comment below.
Has Elon solved the bots problem yet? I haven’t been following too closely. When social media companies are paid by advertisers for ‘page views’, there is no incentive to eliminate bots. Politicians will not get rid of them either because they use bots in campaigns.
Excellent piece, UD. I enjoyed the test.