It is becoming increasingly difficult to distinguish the human face through artificial intelligence. The optics of these imaginary portraits reach levels of completeness, meaning that the viewer considers them to be more reliable and trustworthy than the actual boils. Anyone. If entertaining power, this special effects technology, now accessible and democratized, can be deceptive.
You will be interested too
[EN VIDÉO] Mankind shares 70% of its facial expressions That would be another common feature of the approximately seven billion people who inhabit the earth. Finally, 16 common points. They will share 16 facial expressions in identical contexts.
With Artificial intelligence, It is not easy to distinguish falsehood from true. Especially during this time Technology Used to create intense realistic faces. Researchers at the University of Lancaster have found that it is very difficult to distinguish a real face from something newly created by algorithms.
They did a lot of experiments to find out how realistic these artificial faces are. The researchers first asked 315 participants to identify the faces of the real people and the creators in 800 portraits.Artificial intelligence. They received an accuracy rate of 48.2%, which was slightly lower than the chance. The second group of participants were trained to recognize artificial faces before being subjected to the same experiment. Their results? 59%
These normal scores are not surprising given the level of sophistication of the softwareArtificial intelligence Used to create ultra-realistic faces in just a few clicks. They are available to internet users to the extent that they use them for various purposes Entertainment Or parody. One of them made it appear and watched the fun Elon Musk In the Stanley Kubrick film, 2001, A Space OdysseyOne created by another Deep fake Barack Obama called Donald Trump an “ass”.
This type of display generates a lot of ink and raises many questions about the use of deepfacs to spread Fake news. Researchers at the University of Lancaster wanted to verify that Faces Integration instills confidence in the audience. They asked 223 participants for a score of 1 to 7 to assess the reliability of real and artificial faces. Faces created by artificial intelligence are, on average, considered to be 7.7% more trustworthy than humans. A confusing phenomenon that researchers explain by the fact that Faces The set is very lambda. This overt nature evokes a sense of confidence in most individuals.
“ Our rating Realism The photo of the faces created by the artificial intelligence shows that the set machines have succeeded in crossing the strange valley. [une théorie du roboticien Masahiro Mori selon laquelle l’affinité pour les robots augmente en fonction de leur ressemblance à l’humain, ndlr] And can create indistinguishable – and more reliable – faces than real faces “, Can we read in a recent study published in the journal Science? PNAS.
So be careful when you encounter a photo or A video On the Internet. Especially those created withApplication Chinese Zhao. Its allows users to insert their face inside in a matter of seconds Film juices Or clips with a simple selfie. The result is sometimes impressive realism. On the Internet, seemingly cautious evidence is more important than ever.
Are you interested in reading now?