Wednesday, June 17, 2020

Twitter system to identify 5G coronavirus conspiracy tweets goes haywire

A new system designed to limit coronavirus misinformation has suffered constant errors, drawing scorn and mockery from users

One month after promising to impose warning labels on misleading tweets about coronavirus, Twitter is still failing to accurately distinguish factual information from fake news.

Starting in early May, the company has been using artificial intelligence (AI) to identify misinformation and apply mild or strong warnings depending on how "dangerous" it is, aiming to expand them to other topics in future.

The policy is part of an ongoing attempt to clean up Twitter's service, which has already entangled both it and its rivals Facebook and Snapchat in a furious row with US President Donald Trump.

But since then Twitter's misinformation labels have frequently lapsed into farce, wrongly catching obscene jokes, factual news headlines and complaints about false claims while leaving die-hard conspiracy theorists untouched....

SOURCE 


3 comments:

Anonymous said...

Big Tech has no business deciding what is truth and what isn't.

Stan B said...

It's impossible to make a system fool proof, because fools are so darn creative!

Bill R. said...

The American government has been lying to us since the beginning of this. Is Twitter going to do something about that, too?