Last Week in AI AI for Shaming Politicians, the New AI Art Scene, DeepFake Phishing
Jul 16, 2021
Innovative AI projects are putting politicians in the hot seat with humorous art that critiques their inattentiveness. The explosion of AI-generated art showcases groundbreaking techniques blending text and visuals, while gender bias in machine translation raises industry alarms. A dark twist reveals how deepfakes are weaponized for phishing scams. Meanwhile, Elon Musk grapples with the challenges of self-driving technology as critiques pile up. The episode also highlights advancements in robotics and calls for better oversight of YouTube's recommendation algorithms.
AI Snips
Chapters
Transcript
Episode notes
AI Politician Shaming
- A Belgian artist shames politicians caught using phones at work.
- The AI project uses video feeds and tags phone-using politicians on social media.
AI Art Explosion
- Combining VQGAN and CLIP allows for AI-generated art from text prompts.
- This has led to an explosion of surreal and creative imagery online.
Gender Bias Dataset
- Google AI released a dataset for studying gender bias in machine translation.
- This dataset helps analyze errors like incorrect gender choices and pronoun drops.