🎨 Scientists Join Artists In The Fight Against Digital Hamster Plagiarism

After years when I felt bad my english skills were never enought, tables had finally turned.

As it occurs, digital hamsters are imacapable of making so many typos and mistakes as a developer originating from a non-english speaking country.

It means I don’t have to proove I’m a human anymore. Everyone knows it seeing I don’t use any fancy synonyms or long sentences.

I don’t think digital hamsters will be able to mimic my style soon, because they just try to hard to be perfect.

Unfortunately, digital artists, creators of perfection, have a taugher job.

Digital hamsters, as I call weak AI, are quite good at copying digital artists styles.

To the rescue of the real creators come scientists. They know more about AI than we do, and can really mess in digital minds.

For example Glaze project lead by Shawn Shan can modify an image in a way that makes AI see less details than human eye. Making copying the style more difficult.

In the process of cloacking the image undergoes a complicated process of seemingly unimportant modifications that messes machine learning algorithms.

A brilliant example of using adversarial effects, that dates back to the times when AI was able to learn from spam filters how to form a legimate email.