This weekend in Generative Media

Can the law catch up to AI? Open source pros & cons; Zuck uses Instagram to train AI

• Lawmakers propose anti-nonconsensual AI porn bill after Taylor Swift controversy (The Verge)

• FCC moves to criminalize most AI-generated robocalls (NBC News)

• Law Enforcement Braces for Flood of Child Sex Abuse Images Generated by A.I. (New York Times)

• Finance worker pays out $25 million after video call with deepfake ‘chief financial officer’ (CNN)

• Should we make our most powerful AI models open source to all? (Vox)

• AI2 open sources text-generating AI models — and the data used to train them (TechCrunch)

• Zuckerberg’s Going to Use Your Instagram Photos to Train His AI Machines (PetaPixel)

• Which entertainment jobs are most likely to be disrupted by AI? New study has answers (Los Angeles Times)

• GenAI Can Help Small Companies Level the Playing Field (Harvard Business Review)

• How audio-jacking using gen AI can distort live audio transactions (VentureBeat)

• ‘There is no such thing as a real picture,’ says Samsung exec (The Verge)

• How A.I. Is Remodeling the Fantasy Home (New York Times)

• The Case of the AI George Carlin Comedy Podcast (Vulture)

• AI models are coming to fashion to promote diversity—but some industry insiders are concerned it will end up ‘parodying it’ (Fortune, paywalled)

• Meta is bucking just about every AI trend, including the ‘boys club’ (Semafor)

• The Battle for Better, Broader, More Inclusive AI (IEEE Spectrum)

• An open letter to the artist and songwriter community - why we must call time out on TikTok (Universal Music Group)

• “More than just Auto-Tune on steroids”: researchers claim to have come up with a new AI-powered method of pitch correction that “surpasses previous tools” (MusicRadar)

• Humans prefer AI-generated copy, survey finds (Search Engine Land)

• Microsoft Teams Can Now “Clean Up” Your Webcam Background With AI (HowToGeek)

• AToM: Amortized Text-to-Mesh using 2D Diffusion (project page)

• AnimateLCM: Accelerating the Animation of Personalized Diffusion Models and Adapters with Decoupled Consistency Learning (project page)

• Motion-I2V: Consistent and Controllable Image-to-Video Generation with Explicit Motion Modeling (project page)

• Anything in Any Scene: Photorealistic Video Object Insertion (project page)

• MotionDirector: Motion Customization of Text-to-Video Diffusion Models (API on Replicate)

• How to make a 3D animated character with Charmed (YouTube)

• the new 'style reference' feature in midjourney matches aesthetics so good (X)

• Headshot -> Fullbody -> Video Scene (X)