Back in 1974, a fellow student took time off his dog-killing experiments to code a program for writing porn. It made amusing sentences using a randomised selection of phrases, e.g. "Painfully they peeled a grape for twenty minutes."
Now, there's an AI program that can take an opening line and develop a fairly convincingly human-sounding post from there. Just imagine what that might do if let loose on Facebook, Twitter etc.
https://www.sciencealert.com/scientists-developed-an-ai-so-advanced-they-say-it-s-too-dangerous-to-release
I should think disinformation agencies are watching with keen interest. It could multiply the output of the 77th Brigade, for example:
https://www.wired.co.uk/article/inside-the-77th-brigade-britains-information-warfare-military
Authentication is going to become a big issue.
No comments:
Post a Comment