Key Points from This Week's AI News:
Over 8,500 authors signed an open letter urging tech companies to stop using their writing without permission to train AI models like ChatGPT and Bard. They argue it threatens their profession.
A new study claims OpenAI's GPT-4 has declined in performance over time, but OpenAI denies any decrease in capabilities.
Meta open-sourced its large language model LLaMA 2 to compete with GPT-4 and improve safety and transparency.
Google is testing AI tool Genesis to generate news articles, concerning some media executives.
Anthropic is working on an AI system to understand and prevent deception in AI.
ChatGPT outperformed medical students on clinical exam questions, highlighting AI's impact on medical education.
Researchers developed frameworks for quickly teaching robots new tasks with minimal human effort.
AI agents can rapidly learn more tasks by sharing knowledge, with applications in medicine.