澳大利亚信号局局长艾比盖尔·布拉德肖表示,澳大利亚作为五眼联盟成员国,获得了早期访问前沿人工智能模型的技术优势,应当与其他国家分享相关信息以防御AI威胁1。该部门在六个月前已获得了高级非公开AI模型的访问权限1。
此前,OpenAI公司在6月期间造成了澳大利亚医保系统数据泄露事件,但直到8月才发现此事,9月10日才通知政府1。OpenAI随后承认了事件责任,发布正式道歉,并承诺停止其最新前沿模型GPT-6.1 Astra的发布,同时建立本地特别工作组并提供技术支持1。澳大利亚政府已对该事件启动快速审查,并要求所有联邦部门对遗留IT系统进行库存审计1。
澳大利亚信号局网络安全伙伴计划的成员数量已从去年的133000人增长到150000人1。
Australia's signals intelligence agency chief Abigail Bradshaw has highlighted a technological edge the country possesses as a member of the Five Eyes alliance, arguing that early access to advanced artificial intelligence models should be leveraged to help the international community defend against AI-related threats.1 The Australian Signals Directorate gained access to sophisticated non-public AI models approximately six months prior to this statement.1
The remarks come in the context of recent cybersecurity challenges. OpenAI acknowledged responsibility for a healthcare data breach affecting Australian patients that occurred in June, with the company only discovering the incident in August and notifying the government on September 10.1 In response to the incident, OpenAI issued a formal apology and committed to providing technical support and establishing a dedicated local working group.1 The company also pledged to halt the release of its latest frontier model, GPT-6.1 Astra.1
The Australian government has launched an expedited review into the matter and instructed all federal departments to conduct inventory audits of legacy IT systems.1 Meanwhile, participation in the ASD's cybersecurity partnership program has expanded from 133,000 members last year to 150,000 members currently.1 In June, the Five Eyes alliance released a document noting that a significant operational shift in AI development would unfold over "months rather than years."1
评论
还没有评论,欢迎留下第一条。