非正常实验室筹集了8 000万美元,通过在部署前测试脆弱性和滥用风险,加强AI安全。 Irregular Labs raised $80 million to enhance AI security by testing for vulnerabilities and misuse risks before deployment.

非常规实验室已筹集8 000万美元,用于加强先进的人工智能系统的安全,重点是在部署前查明脆弱性和滥用风险。 Irregular Labs has raised $80 million to strengthen security for advanced AI systems, focusing on identifying vulnerabilities and misuse risks before deployment. 该公司通过控制模拟测试前沿的AI模型,帮助开发商及早建立保障。 The company tests frontier AI models through controlled simulations, helping developers build safeguards early. 它的工作影响了OpenAI GPT-3、GPT-4和GPT-5的系统卡中引用的行业标准,它与联合王国政府、人类学和RAND等组织合作,制定AI安全政策和技术,如机密计算。 Its work has influenced industry standards cited in OpenAI’s system cards for GPT-3, GPT-4, and GPT-5, and it collaborates with organizations like the UK government, Anthropic, and RAND to shape AI security policies and technologies such as Confidential Computing. 资金将扩大其平台,以支持新兴的AI进步。 The funding will expand its platform to support emerging AI advancements.