OpenAI is making significant advancements in AI technology with the formation of a Safety and Security committee and the training of their next Frontier Model towards achieving Artificial General Intelligence (AGI). Google and Microsoft are also progressing in the AI space, with Google updating its Gemini 1.5 Pro API and Microsoft facing privacy concerns over its Recoil AI feature, while Google’s search generative experience has received criticism for providing inaccurate responses.
In recent news from OpenAI, a new Safety and Security committee has been formed to make recommendations on critical decisions for all projects within the company. The board has also announced that they have started training their next Frontier Model, hinting at the development of a new AI system with capabilities beyond what is currently anticipated. This new model is expected to be a significant step towards achieving Artificial General Intelligence (AGI), and the wording used suggests a groundbreaking advancement in AI technology.
The departure of members from OpenAI’s superintelligence team has shed light on the company’s focus on AGI rather than superintelligence. This shift is exemplified by the development of new AI models and the training of GPT-5, which has undergone red teaming to ensure its effectiveness and safety. Additionally, the company’s collaboration with Microsoft hints at the imminent release of a new model, possibly named differently from GPT-5, which is expected to offer enhanced capabilities.
Google has updated its Gemini 1.5 Pro API through post-training to improve the model’s performance, showcasing advancements in AI technology. The competition among top AI models such as GPT-4, Gemini Advanced, and others is heating up as they strive to enhance their capabilities and outperform each other in various metrics. The convergence of these models indicates rapid advancements in the AI space and raises anticipation for future developments in the field.
Microsoft’s new Recoil AI feature, which tracks users’ computer activities, has sparked privacy concerns and is currently under investigation. The feature allows users to review their computer usage history, raising questions about data security and privacy implications. While Recoil offers potential benefits in terms of data organization and retrieval, its data tracking capabilities have raised red flags regarding user privacy and information security.
In another development, Google’s search generative experience has faced criticism for providing inaccurate and sometimes humorous responses sourced from various websites, including Reddit. The feature aims to offer concise answers to user queries but has received backlash for delivering misleading information. Users are advised to exercise caution when using this feature to avoid relying on potentially inaccurate or misleading responses.