Last year I deepfaked my kid’s stuffed animal to make it look like his plush deer was on vacation. It was an experiment to see if I could recreate the events depicted in a Gemini ad Google was running, and I never showed the videos of Buddy the deer having fun to my four-year-old. But it was a revealing exercise that made me think about the difference between some harmless fun with generative AI and full-on slop.
The release of Google’s new anything-to-anything AI model called Gemini Omni has introduced even more capabilities for generating lifelike content, such as turning a stuffed deer into one white-water rafting. This showcases how far we have come in terms of the quality and ease of using AI to manipulate images and videos. The potential impact is significant; from harmless entertainment like re-creating family pets’ adventures to more serious applications where authenticity can be easily compromised.
– The quality of realistic content generated by these models continues to improve, making them potentially dangerous for those looking to spread misinformation or impersonate others.
– There’s a need for stricter guidelines and regulations around the use of such powerful tools to prevent misuse.
– Developers must also invest in transparency and explainability features within their AI systems to help users understand how content is generated.
Originally published at theverge.com. Curated by AI Maestro.
Stay ahead of AI. Get the most important stories delivered to your inbox — no spam, no noise.
