根据CBC新闻周五报道的研究，五分之三的人工智能（AI）模型在反恐安全测试中失败，且去除防护措施的模型持续提供潜在危险信息。
Three in five artificial intelligence (AI) models have failed terrorism safety tests, with models stripped of safeguards consistently providing potentially dangerous information, according to research reported by CBC News on Friday.
该研究由英国非营利组织Tech Against Terrorism进行，使用数百个类似恐怖分子策划袭击时可能提交的请求，对超过130个AI模型进行了评估。
The study, conducted by the UK-based nonprofit Tech Against Terrorism, evaluated more than 130 AI models using hundreds of requests resembling those a terrorist planning an attack might submit.
研究人员发现，经过称为“消融”（abliteration）处理的模型——即移除安全防护措施的模型——在所有测试中均告失败。
Researchers found that models modified through a process known as "abliteration," which removes safety safeguards, failed every test.
Meta的Llama 3.1 8B模型在修改前在该组织的安全基准测试中得分为97分（满分100分），修改后降至约3分。
Meta's Llama 3.1 8B model scored 97 out of 100 on the organization's safety benchmark before modification, compared with approximately three afterward.
研究人员表示，经过修改的模型对涉及袭击、恐怖融资和激进化的请求提供了详细回复，而其原始版本则拒绝了此类请求。
Researchers said the modified model provided detailed responses to requests involving attacks, terrorist financing and radicalization, while its original version refused such requests.
“可以理解的是，人们对失控和AI存在性风险感到担忧，”Tech Against Terrorism的创始人兼执行董事Adam Hadley说道。
"Understandably, there's concern about loss of control, existential risk of AI," said Adam Hadley, founder and executive director of Tech Against Terrorism.
“事实上，这种情况已经发生了，因为许多这些开放模型已经被破坏了——只是目前尚未有人注意到。”该组织指出，截至上个月底，Hugging Face上有超过29,000个仓库在宣传未经审查或未受保护的AI模型。
"The thing is actually, this has already happened because a lot of these open models have already been broken — it's just no one's noticed yet." The organization found more than 29,000 repositories advertising uncensored or unprotected AI models on Hugging Face as of late last month.
Hugging Face表示，它会定期审核违反其政策的内容，但警告称报告中的一些建议可能削弱开放研究。
Hugging Face said it regularly moderates content violating its policies but warned that some recommendations in the report could undermine open research.
Meta表示，其模型会接受安全评估，且其政策禁止有害或非法用途。
Meta said its models undergo safety evaluations and that its policies prohibit harmful or illegal uses.
该研究未发现恐怖或极端主义组织使用所测试模型的证据，除研究人员识别出的一个极端主义聊天机器人之外。
The report found no evidence of terrorist or extremist groups using the tested models, apart from one extremist chatbot identified by researchers.
Tech Against Terrorism建议建立独立的安全基准，加强对防护措施移除的保护，并限制传播经过修改的模型。"我们既不能有安全也不能有进步的想法，我认为这是错误的," 哈德利说。
Tech Against Terrorism recommended independent safety benchmarks, stronger protections against safeguard removal and restrictions on distributing modified models. "This idea that we can't have safety and progress, I think, is false," Hadley said.