YouTube lecture here

As part of Fast.ai’s NLP course, Rachel Thomas has prepared an excellent overview of the existing cases of disinformation. She highlights several notable examples from this past year alone, and covers the potential misuses and abuses of AI and NLP techniques. These include prominent generative methods such as GPT-2 to make fake-but-plausible reddit comments, and StyleGAN portraits to fake social media profile images.