News for hackers. Stuff that matters.
Re:Grover – A State-of-the-Art Defense Against Neural Fake News by jasonhansel on Saturday June 08, 2019 @05:33PM (#20135324)
(1) Generate 100 fake news articles

(2) Remove the 92 articles that Grover detects (with 92% accuracy rate)

(3) Choose the best of the remaining 8 articles

Re:Grover – A State-of-the-Art Defense Against Neural Fake News by hippiecow on Saturday June 08, 2019 @05:54PM (#20135417)
"Note that, even if Grover fails to detect a given piece as fake, our findings suggest that releasing many such articles taken together would be relatively easy to spot. Thus, if a source of Neural Fake News disseminates a large number of articles, Grover will be increasingly capable of spotting these articles as malicious."
Re:Grover – A State-of-the-Art Defense Against Neural Fake News by hamburglar1 on Saturday June 08, 2019 @06:27PM (#20135616)
That doesn't solve the exploit above
Re:Grover – A State-of-the-Art Defense Against Neural Fake News by hippiecow on Saturday June 08, 2019 @06:55PM (#20135755)
My interpretation is that it would, since if an adversary released a lot of rejection-sampled articles and you retrained Grover on them, then Grover could tag them all as machine-generated.
Re:Grover – A State-of-the-Art Defense Against Neural Fake News by jasonhansel on Saturday June 08, 2019 @07:48PM (#20135975)
Of course, to retrain Grover, you'd have to already know that they were machine-generated.
Re:Grover – A State-of-the-Art Defense Against Neural Fake News by steve19 on Sunday June 09, 2019 @12:43AM (#20137172)
The generated news also needs to look like it was written by a human and be readable by the average human.

So you need a step that filters out obvious junk, maybe by eyeballing or with another net.

It is possible that the only way to beat grover is rubbish such as contorted, but valid, grammer or overuse of synonyms.