Distinction Between Human and Synthetic Content is Challenging, Said 69.74%

According to a recent announcement, Meta will label AI-generated content among its platforms. “In the coming months, we will label images that users post to Facebook, Instagram, and Threads when we can detect industry-standard indicators that they are AI-generated,” stated Meta.

Real Research conducted a survey on Meta labeling AI-generated content and found that 69.74% of the respondents agreed with this initiative and 30.26% did not.  

Highlights:

• 72.08% believe labeling AI images could set human content creators apart and create a higher demand for them.
• The distinction between human and AI content becomes increasingly challenging to discern, said 69.74%.
• 55.88% said AI content can alter people’s perspectives when voting.

Meta Will Label AI-Generated Content

While 72.08% believe that labeling AI images could set human content creators apart and create a higher demand for them, 27.92% don’t agree. As per the announcement, Meta will label AI-generated content as the distinction between human and synthetic content increasingly becomes challenging to discern. 69.74% agreed with this sentiment, and 30.26% did not.

What Could Go Wrong?

Senior Executive at Meta, Sir Nick Clegg, stated to Reuters that the technology is “not yet fully mature.” After the conclusion that Meta will label AI-generated content, Real Research surveyed this judgment and found that 69.88% think actors may take advantage of this new technology and present AI content as human-created. 30.12% did not agree with this sentiment.

Concerns about Meta Technology

It is a given that Meta will label AI-generated content; this is what Professor Soheil Feiz, director of the reliable AI lab at the University of Maryland, said about the initiative. The professor said, “Although Meta would train our detectors, it could be easily evaded by some lightweight processing on top of the images.” 49.6% are somewhat concerned about this, 39.92% are not concerned at all, and 10.48% are extremely concerned.

Meta to Label AI-Generated Images and Videos?

Meta announced that its tool will not work on audio and video content, while there are widespread concerns about fake video and audio content. 50.04% believe it is important to label AI-generated fake video and audio content; 44.04% said it is very important; and 5.94% said it was not important at all. 

Meta requested that users label the audio and videos that they publish; it stated that those who fail to label could face penalties. 54.34% of respondents are somewhat likely to follow this protocol, and 34.76% are very likely to follow it. 10.90% said that it is not likely that people will follow this protocol.  

Meta stated that they are taking this approach through the next year, during which several important elections are taking place around the world. 55.88% agree that AI content can alter people’s perspectives when it comes to voting, and 39.14% somewhat agree. 4.98% did not agree with this sentiment at all.