A new generation of artificial intelligence (AI) models can produce “creative” images on-demand based on a text prompt.
By prompting DALL-E 2 to create images containing text captions, then feeding the resulting (gibberish) captions back into the system, the researchers concluded DALL-E 2 thinks Vicootes means “vegetables”, while Wa ch zod rea refers to “sea creatures that a whale might eat”.
One possibility is the “gibberish” phrases are related to words from non-English languages. For example, Apoploe, which seems to create images of birds, is like the Latin Apodidae, which is the binomial name of a family of bird species. This seems like a logical explanation.
One point that supports this theory is the fact that AI language models don’t read the text the way humans do. Instead, they break input text up into “tokens” before processing it.
Different “tokenization” approaches have different results. Treating each word as a token seems like an intuitive approach but causes trouble when identical tokens have different meanings (like how “match” means different things when you’re playing tennis and when you’re starting a fire).
Contrarily, treating each character as a token produces a smaller number of possible tokens, but each one conveys much less meaningful information.
The “secret language” could also just be an example of the “garbage in, garbage out” principle. DALL-E 2 can’t say “I don’t know what you’re talking about”, so it will always generate an image from the given input text.
DALL-E’s “secret language” is an example of an “adversarial attack” against a machine learning system: a way to break the intended behavior of the system by intentionally choosing inputs the AI doesn’t handle well.
Adversarial attacks raise security concerns. DALL-E 2 filters input text to stop users from producing harmful or abusive content, but a “secret language” of gibberish words might enable users to bypass these filters.
There is a freely available smaller model, DALL-E mini can be used to check it out too!