Weekly AI News - Jun. 20, 2025

This week on our show, the hosts discuss Google's Gemini transparency cut hindering enterprise developers. They examine Grok's responses highlighting AI alignment issues, and MIT researchers developing SEAL, an AI model for continuous learning via synthetic data. They explore DeepSeek claiming its R1 model matches Google/Anthropic coding ability, and the African Development Bank with Google exploring strategies for Africa's AI future. The show then reviews Nvidia's 'sovereign AI' strategy gaining traction among EU leaders, and 'The OpenAI Files' report detailing OpenAI's evolution and governance concerns. They also look at Pope Leo XIV making AI's threat to humanity a signature issue, and Amazon CEO Andy Jassy expecting AI to shrink the company's workforce. Finally, they mention Geoffrey Hinton suggesting manual dexterity jobs are safest from AI, predicting assistants will be replaced. Hosted on Acast. See acast.com/privacy for more information.