If you are an avid social media user, you probably have come across a viral post in which a Twitter user indicated that he had managed to force an artificial intelligence to see 1,000 hours of ads from Olive Garden, an Italian-American restaurant chain. from the United States, and then force her to write her own publicity.
Although its absurd context may have led some users to laugh out loud, the truth is that the publication was fake, as were all the other bot adventures published by Keaton Patti, an American comedian. As the Gizmodo website demystifies, artificial neural networks are intelligent, but not yet able to do so, let alone understand the nuances of satire.
However, this type of artificial intelligence manages to accomplish some interesting tasks in its own way. According to Andrej Karpathy, expert in the field and director of artificial intelligence and autopilot vision at Tesla, a simple artificial neural network can generate text versions of Shakespeare, scientific articles, and even Wikipedia.
Inspired not only by this premise, but also by Keaton Patti's viral tweets, Jonathan Fly, a machine learning and artificial intelligence recognition and understanding enthusiast, decided to create the I Forced a bot… site to let the world know what happens when a bot is required to perform certain actions.
Aimed at anyone interested in how artificial neural networks work, as well as those just curious, I Forced a bot presents itself as a collection of real experiences ranging from caricature to usefulness.
Johnathan Fly's latest adventure, where he trained an artificial intelligence used to reconstruct facial images in an experience with emoji and video game characters with terrifying results, has even caught the attention of the Motherboard website.