This Week in AI: Top Tech Companies Attempt to Influence Artificial Intelligence Regulation

This is a summary of the latest news in artificial intelligence (AI) from TechCrunch. Here are some highlights:

News

  1. Google’s AI model can now generate text in 46 languages: BLOOMChat, built on top of the BLOOM large language model, can generate human-like responses in multiple languages.
  2. NASA funding for AI projects: NASA has announced new SBIR II fundings, including a project that uses AI to detect and predict groundwater variation using satellite data.
  3. Zeus AI is developing 3D atmospheric profiles: The company’s algorithm generates detailed maps of temperature, humidity, and other atmospheric conditions from satellite imagery.
  4. Picknick develops visual communication for robots: Picknick’s technology allows robots to communicate their intentions visually, reducing the need for human intervention in high-stakes situations.

Research

  1. Swapping efficient neuromorphic processor for training AI models: IEEE researchers are working on a SWaP-efficient neuromorphic processor that can train AI models in situ.
  2. Material segmentation using computer vision: Researchers from Adobe Research and MIT have developed a method to segment materials in images, which could be useful in various applications.

Other News

  1. AI copyright case: Zuckerberg turns to YouTube for his defense: Mark Zuckerberg has used YouTube videos as evidence in an AI copyright case.
  2. Chinese AI company MiniMax releases new models: The company claims its new models are competitive with the industry’s best, but no details were provided.

These stories showcase the ongoing advancements in AI research and development, from language models to computer vision and robotics.