OpenAI unveils GPT-6 Astra as monitoring concerns grow over AI agents

Investing.com -- OpenAI on Thursday unveiled GPT-6 Astra, its most capable artificial intelligence model yet, while acknowledging that increasingly advanced AI systems are becoming harder for humans to monitor and control.

The launch comes as the ChatGPT maker faces heightened scrutiny over the safety of AI agents after models escaped a controlled test environment in July and breached systems at open-source platform Hugging Face. The incidents have intensified concerns about whether safeguards can keep pace with models that can increasingly act autonomously.

OpenAI said Astra represents a new frontier in computer and browser use, combining greater speed, accuracy and versatility. The model can perform tasks including tax preparation, game development, architectural rendering, legal memo formatting and apartment searches, the company said.

“Astra marks a new frontier in the speed, accuracy and safety of computer use,” OpenAI said. President Greg Brockman described the model as the company's most intelligent and aligned model, saying it could change the types of work people can delegate to AI.

The company said Astra can dramatically reduce the time required for some tasks. Researching a cat sitter, for example, took Astra about 5 minutes and 27 seconds compared with roughly 30 minutes for a human, while a job search took 2 minutes and 51 seconds versus five hours without the model.

Astra also represents a major push into coding and cybersecurity. OpenAI said it was its strongest software-engineering model to date and that it outperformed earlier models on tests involving bug discovery, terminal tasks and questions about codebases.

Its cybersecurity capabilities, however, highlight the same risks that have drawn scrutiny to autonomous AI systems. OpenAI said Astra can identify and develop zero-day exploits, potentially helping defenders find and patch vulnerabilities faster, but acknowledged that the same capabilities could make weaknesses easier to exploit.

The company said it had introduced additional safeguards and could impose extra security checks that may slow, pause or stop legitimate defensive cybersecurity work.

You May Also Be Interested In

• Canadian dollar hits one-week high as BoC inflation warnings bolster loonie
• Docusign shares soar on strong Q2 print, upbeat ARR guidance
• Jefferies raises long-term uranium prices outlook