OpenAI每天处理50拍字节的数据，如今安全方面的投入已占计算成本10%以上。澳大利亚悉尼——OpenAI的一位高级主管称，该公司未经授权获取澳大利亚医保统计数据的行为“并非特别复杂”，并呼吁更多地运用人工智能来强化网络安全，以应对这一新兴技术带来的风险。
OpenAI reviews 50 petabytes of data, safety now costs more than 10% of compute SYDNEY, Australia — A senior OpenAI executive has described its unauthorized access of Medicare statistical data in Australia as “not super sophisticated,” encouraging greater use of AI to shore up cybersecurity defenses as the antidote to risks from the emerging technology.
战略总监杰森·权在周二出席澳大利亚议会人工智能特别联合委员会听证会时表示，OpenAI支持强制性的安全事件上报制度，并再次就未经授权访问澳大利亚政府网站一事道歉。
Chief strategy officer Jason Kwon told an Australian parliamentary inquiry OpenAI supports mandatory incident reporting and reiterated an apology for unauthorized access to Australian government websites in his appearance at the Joint Select Committee on Artificial Intelligence on Tuesday.
今年9月，澳大利亚总理安东尼·阿尔巴尼斯在纽约呼吁国际社会共同制定人工智能最低标准时披露了这起事件。此后不久，OpenAI也公布了全球范围内发生的多起类似事件。在澳大利亚，舆论批评的焦点在于：OpenAI在发现有人未经授权访问澳大利亚服务局网站后，竟拖延了一个月才通知该机构。目前，OpenAI已向受影响的澳大利亚政府机构及私营合作伙伴开放其斥资10亿美元设立的“黎明行动”基金。面对反对党联盟的质询，杰森·权仍将此次入侵行为形容为“并非特别复杂”。
Australian Prime Minister Anthony Albanese revealed the incident in New York in September while rallying international support for minimum standards on AI, shortly before OpenAI disclosed a string of such incidents worldwide. Criticism in Australia focused on a one-month delay between OpenAI detecting unauthorized access to a Services Australia website and its decision to inform the agency via a generic OpenAI has offered affected Australian government agencies and private partners access to its $1 billion Operation Daybreak fund. Under questioning from the Coalition Opposition, Kwon described the breach as “not super sophisticated.”