Weekly bulletin from AIport, issue #8
The UN adopts its first AI resolution, France fines Google €250 million, Sakana releases a model-merging solution, and much more.
Hello and welcome to issue #8 of our weekly bulletin! We hope you enjoy this week’s read. Don’t forget to subscribe, comment, and reach out to us with feedback or new suggestions.
North America:
In New York, the UN General Assembly adopts its first AI resolution that safeguards user privacy and human rights. In Nashville, Tennessee Governor Bill Lee signs the ELVIS Act to protect the state’s artists from AI deepfakes.
GPT-5 from OpenAI will be released this summer, claims Business Insider.
DeepMind unveils TacticAI, a predictive GenAI solution that provides insights to soccer coaches, particularly on corner kicks.
YouTube updates its guidelines to require special labels for AI-generated videos.
Elon Musk’s xAI releases the weights and architecture of its Mixture-of-Experts model, Grok-1.
Europe:
In Denmark, a man is sentenced to prison for conducting a €270K music-streaming fraud powered by AI.
France’s Competition Authority fines Google €250 million for training its AI chatbot Bard (now Gemini) on content from third parties without their consent.
Asia:
Tokyo-based Sakana AI rolls out Evolutionary Model Merge, a methodology that automatically combines multiple foundation models into novel solutions with enhanced AI capabilities.
Saudi Arabia plans to launch a $40 billion fund specifically for AI projects, reports the New York Times.
In China, a tech-savvy grandson uses AI to impersonate his recently deceased father, shielding his oblivious grandmother from grief.
Africa:
An NLP research group founder from Ghana focuses his efforts on preserving African languages.