developing · Story package
OpenAI classifies upcoming Astra model as cybersecurity-critical
The company says it is applying additional controls under its Preparedness Framework, but has not supplied an independent capability evaluation or detailed control specification.
Overview
OpenAI says Astra, an upcoming model, is the first model it is handling as critical for cybersecurity under its Preparedness Framework.
The company says additional controls are being applied during development and frames the effort around making advanced cyber capabilities available to defenders. The bounded source does not provide an independent capability evaluation or detailed controls.
Why it matters
A critical designation signals a higher claimed cybersecurity capability and risk tier than OpenAI has previously assigned to a model.
The control design and release conditions will affect both defensive access and misuse risk.
Key facts
OpenAI says Astra is its first cybersecurity-critical model under the company's Preparedness Framework.
OpenAI says additional controls are being applied while Astra is in development.
The source does not include an independent capability evaluation or a detailed control specification.
Latest update
No public update is available.
Full timeline
No public timeline entries are available.
Sources
Open questions
- What capabilities triggered the critical classification?
- What controls and release restrictions will apply?
- When and in what form will Astra become available?
Related stories
No related stories are listed.
Roundup appearances
No roundup appearances are available.