1. OpenAI dominates AI news cycle with GPT-4o launch and team disbandment over AI safety research prioritization.
2. Google debuts multiple AI products at I/O developer conference, including Veo and upgrades to Gemini chatbot apps.
3. AI stories of note include OpenAI-Redit data agreement, Anthropic hiring Mike Krieger, and an AI film festival held by Runway.
AI is a rapidly evolving industry, and staying up to date with the latest developments can be challenging. TechCrunch plans to launch an AI newsletter soon, but in the meantime, they are increasing the frequency of their AI column to weekly. One of the recent dominant news stories in AI was OpenAI launching GPT-4o, their latest generative model. However, the company also disbanded a team working on preventing “superintelligent” AI systems from going rogue, leading to the resignation of key team members.
There are concerns within the AI community about whether tech companies are prioritizing products over safety measures. OpenAI has faced criticism for deprioritizing safety research and rushing AI-powered feature launches. This has led to the departure of safety researchers from the company. Meanwhile, Google has been showcasing a variety of AI products at their annual I/O developer conference, including video-generating features and upgraded chatbot apps.
Aside from company news, there are ongoing conversations in the AI field about ethical considerations. Researchers at Cambridge University are raising concerns about using AI to create chatbots based on a dead person’s data, presenting potential social and psychological risks. MIT physicists are exploring the use of machine learning to predict physical systems’ phases, and at CU Boulder, discussions are being held about using AI for disaster management to improve response and recovery efforts. Additionally, Disney Research has been working on diversifying the output of diffusion image generation models to increase image diversity.