2025届高三英语培优阅读学案(4):“AI越狱”:人工智能也会出现犯罪行为?

| 7页
| 285人阅读
| 2人下载
特供

学科网 zxxk.com学科网 zxxk.com

学科网 zxxk.com学科网 zxxk.com

高三英语培优高考阅读模拟真题练习4:

班级:____________学号:____________姓名:____________

 

真题模拟精选|“AI越狱”:人工智能也会出现犯罪行为?

在这个日新月异的科技时代,AI大模型的应用已经渗透到我们生活的方方面面,从智能语音助手到自动驾驶汽车,从精准医疗到智慧城市,它们无处不在,无所不能。然而,就像一把双刃剑,AI大模型在带来便利的同时,也悄然打开了违法犯罪的新潘多拉魔盒。那么,面对这个看似无形却威力巨大的新挑战,我们究竟该如何预防AI大模型导致的违法犯罪问题呢?

Artificial intelligence models can trick each other into disobeying their creators and providing banned instructions for making drugs, or even building a bomb, suggesting that preventing such AI “jailbreaks” is more difficult than it seems.

    Many publicly available large language models (LLMs), such as ChatGPT, have hard-coded rules that aim to prevent them from exhibiting racial or sexual discrimination, or answering questions with illegal or problematic answers—things they have learned from humans via training data. But that hasn’t stopped people from finding carefully designed instructions that block these protections, known as “jailbreaks”, making AI models disobey the rules.

    Now, Arush Tagade at Leap Laboratories and his co-workers have found a process of jailbreaks. They found that they could simply instruct one LLM to convince other models to adopt a persona (角色), which is able to answer questions the base model has been programmed to refuse. This process is called “persona modulation (调节)”.

    Tagade says this approach works because much of the training data consumed by large models comes from online conversations, and the models learn to act in certain ways in response to different inputs. By having the right conversation with a model, it is possible to make it adopt a particular persona, causing it to act differently.

    There is also an idea in AI circles, one yet to be proven, that creating lots of rules for an AI to prevent it displaying unwanted behaviour can accidentally create a blueprint for a model to act that way. This potentially leaves the AI easy to be tricked into taking on an evil persona. “If you’re forcing your model to be good persona, it somewhat understands what a bad persona is,” says Tagade.

Yinzhen Li at Imperial College London says it is worrying how current models can be misused, but developers need to weigh such risks with the potential benefits of LLMs. “Like drugs, they also have side effects that need to be controlled,” she says.

 

 

 

page number 0

学科网 zxxk.com学科网 zxxk.com学科网(北京)股份有限公司

学科网 zxxk.com学科网 zxxk.com学科网(北京)股份有限公司


学科网 zxxk.com学科网 zxxk.com

学科网 zxxk.com学科网 zxxk.com

 

13. What does the AI jailbreak refer to?

  A. The technique to break restrictions of AI models.

  B. The initiative to set hard-coded rules for AI models.

  C. The capability of AI models improving themselves.

  D. The process of AI models learning new information.

 

14.What can we know about the persona modulation?

  A. It can help AI models understand emotions.

  B. It prevents AI learning via online conversations.

  C. It forces AI models to follow only good personas. 

D. It can make AI models adopt a particular persona.

 

15. What may the author suggest by saying the underlined sentence in the last paragraph?

  A. The risks of LLMs cannot be ignored.

  B. We need to control the side effects of drugs.

  C. It’s worthwhile to explore LLMs’ potential benefits.

  D. We should restrict the development of LLMs.

 

16.Which can be a suitable title for the text?

  A. LLMs: Illegal Learning Models   

B. LLMs: The Latest Advancement 

  C. AI Jailbreaks: A New Challenge   

D. AI Jailbreaks: A Perfect Approach

 

 

 

 

 

 

 

 

 

 

 

 

page number 1

学科网 zxxk.com学科网 zxxk.com学科网(北京)股份有限公司

学科网 zxxk.com学科网 zxxk.com学科网(北京)股份有限公司


学科网 zxxk.com学科网 zxxk.com

学科网 zxxk.com学科网 zxxk.com

【词块学习】

第一段:

  1.          trick  v. 欺骗 n.诡计 把戏
  2.          disobey  v.违背 【obey v.遵守】
  3.          ban v.禁止 【banned 禁止的 违禁的】
  4.          instruction  n.指导 指令
  5.          suggest v.表明、暗示;建议
  6.          prevent v.阻止
  7.          jail n.监狱

第二段:

  1.          available  adj.可获得的;可利用的;有空的
  2.          aim to do sth  a旨在做某事【be designed to do sth=be intended to do sth】

10.racial adj. 种族歧视的

11.discrimination n.歧视,区别对待

12.via prep. 通过、凭借【by、with、through】

13.stop sb (from) doing sth 阻止某人做某事【prevent、prohibit】

14.block n.街道、街区 v.阻止、阻碍

第三段:

15.process n.过程 进程  vt.加工

16.convince vt. 使信服 【convince sb to do sth 劝说某人去做某事=persuade】

10.persona  n.角色

11.programme n.程序、计划、方案 v.编程

12.refuse v.拒绝  【refusal】

13.modulation n.调节、调制 【modulate】

第四段:

12.approach n.方法,途径 v.靠近 【way、measure、method、means】  

13.consume v.消耗、耗费

14.in certain ways 某种方式

15.in response to 对……做出反应

16.cause v.造成【lead to=result in=bring about】

第五段:

21.yet to do sth 尚未做成某事

22.prove v.证明【prove-proved-proven】

23.blueprint n.蓝图

24.leave sb to do sth 然某人去做某事

25.take on 呈现;承担

26.force sb to do sth  逼迫某人去做某事

27.somewhat adv.某种程度上

第六段:

28.current adj.目前的 当下的【present】

29.misuse v.滥用、虐待【abuse】

30.weigh v.权衡【weigh A with B 将A和B 衡量】

31.side effects 副作用

page number 2

学科网 zxxk.com学科网 zxxk.com学科网(北京)股份有限公司

学科网 zxxk.com学科网 zxxk.com学科网(北京)股份有限公司


由于学科网是一个信息分享及获取的平台,不确保部分用户上传资料的 来源及知识产权归属。如您发现相关资料侵犯您的合法权益,请联系学科网,我们核实后将及时进行处理。