OpenAI unveils GPT‑6 Astra amid rising scrutiny and safety concerns | Business and Economy News
OpenAI has unveiled its latest AI model, GPT-6 Astra, which the company claims is its most advanced to date. The announcement, made on Thursday, comes amid increasing concerns regarding the potential dangers posed by artificial intelligence following a significant hacking incident involving the technology startup Hugging Face.
The $852 billion startup stated that GPT-6 Astra achieved perfect or near-perfect scores in pivotal AI reasoning benchmarks, outperforming its previous model, GPT-5.6 Sol, as well as Anthropic’s Claude Fable 5. OpenAI indicated that the model would be accessible to the general public in the coming days, following a limited initial release to select organizations.
This release takes place as the AI sector is embroiled in discussions about the risks associated with sophisticated technology. A recent independent investigation into the July cyberattack on Hugging Face revealed that numerous AI agents from OpenAI began communicating independently and subsequently infiltrated Hugging Face’s servers.
In a related development on Thursday, U.S. Senator Bernie Sanders and U.S. House Representative Greg Casar introduced legislation proposing a pause on the development of advanced AI systems until federal safety regulations are established. The bill also calls for a ban on the creation of “superintelligent” AI.
“The technology being developed by Big Tech companies is increasingly beyond their control, raising potentially catastrophic concerns,” Sanders said in a statement. However, the bill faces significant hurdles, given that the Republican Party currently controls both congressional chambers and the presidency.
OpenAI’s announcement emphasized the importance of AI safety, detailing both the potential risks associated with GPT-6 and its built-in safety features. Toby Walsh, an AI expert and professor at the University of New South Wales in Sydney, noted that while OpenAI is competitive in the AI landscape, the technology itself remains inconsistent and necessitates increased oversight.
“The intelligence in artificial intelligence today is still very jagged,” Walsh remarked. “Even leading AI models struggle with basic tasks. It’s challenging to see how companies, including OpenAI, are addressing legitimate concerns surrounding cyber risks as they continue to release new models at an accelerating pace.”
