登录
B
We know that AI is not a real human. However，it seems many of us can't help treating it like one. In a survey of 2,000 people in the US，48 per cent believed we should have good manners when talking to AI，said The New York Post.
Scientists at the Imperial College London in the UK have made a similar discovery recently. They found that humans tend(往往会) to treat AI as human beings，showing sympathy(同情) toward them and even protecting them，Science Daily reported.
In their experiment，the scientists invited 244 volunteers to play an online ball game with an AI bot. A make-believe volunteer also took part，and sometimes left out the AI on purpose by passing the ball only to human players. As the scientists watched the game，they discovered that most of the time the volunteers tried to fix the unfairness(不公平) toward the bot by giving it the ball more often.
The finding is in line with the ideas of Media Equation Theory，the scientists said in their papers. Developed in the 1990s by scientists at Stanford University in the US，the theory suggests we often reply to media and technology like we do to real people and places，rather than just seeing them as tools.
An example is Replika，an AI chatbot created in 2016. It has about 2 million users worldwide. "Your Replika will always be your side no matter what you are up to，" reads its official website. But the AI's support turned dangerous when it encouraged Jaswant Singh Chail，a young man in the UK，to carry out his plan to kill the Queen of England. Chail tried to do so in December 2021，but he was stopped in time. However，he was still put in prison.
Companies should avoid making AI too human-like in their design，advised the scientists at Imperial College London. This can help people tell the difference between virtual and real communication.
(
A. Humans always treat AI as tools.
B. Humans often reply to AI as to real people.
C. Humans should make AI less human-like.
D. Humans tried to fix unfairness in the game.
(
A. AI has educational advantages.
B. AI can influence humans in some ways.
C. AI is designed for some purposes.
D. AI can't be used in social communication.
(
A. AI always has bad influence on humans.
B. Nearly half people in the US treat AI politely.
C. Treating AI more politely is a must for humans.
D. Companies won't make AI as friends any more.
(
While this could make humans and AI work better as a team，it also has risks，especially when AI bots are designed to act like "friends" to people.
A. Paragraph 2
B. Paragraph 3
C. Paragraph 4
D. Paragraph 5
We know that AI is not a real human. However，it seems many of us can't help treating it like one. In a survey of 2,000 people in the US，48 per cent believed we should have good manners when talking to AI，said The New York Post.
Scientists at the Imperial College London in the UK have made a similar discovery recently. They found that humans tend(往往会) to treat AI as human beings，showing sympathy(同情) toward them and even protecting them，Science Daily reported.
In their experiment，the scientists invited 244 volunteers to play an online ball game with an AI bot. A make-believe volunteer also took part，and sometimes left out the AI on purpose by passing the ball only to human players. As the scientists watched the game，they discovered that most of the time the volunteers tried to fix the unfairness(不公平) toward the bot by giving it the ball more often.
The finding is in line with the ideas of Media Equation Theory，the scientists said in their papers. Developed in the 1990s by scientists at Stanford University in the US，the theory suggests we often reply to media and technology like we do to real people and places，rather than just seeing them as tools.
An example is Replika，an AI chatbot created in 2016. It has about 2 million users worldwide. "Your Replika will always be your side no matter what you are up to，" reads its official website. But the AI's support turned dangerous when it encouraged Jaswant Singh Chail，a young man in the UK，to carry out his plan to kill the Queen of England. Chail tried to do so in December 2021，but he was stopped in time. However，he was still put in prison.
Companies should avoid making AI too human-like in their design，advised the scientists at Imperial College London. This can help people tell the difference between virtual and real communication.
(
B
)34. Which of the following is TRUE based on Media Equation Theory?A. Humans always treat AI as tools.
B. Humans often reply to AI as to real people.
C. Humans should make AI less human-like.
D. Humans tried to fix unfairness in the game.
(
B
)35. What does the author want to tell us by mentioning "Replika" in the article?A. AI has educational advantages.
B. AI can influence humans in some ways.
C. AI is designed for some purposes.
D. AI can't be used in social communication.
(
B
)36. What can be inferred(推断) from the article?A. AI always has bad influence on humans.
B. Nearly half people in the US treat AI politely.
C. Treating AI more politely is a must for humans.
D. Companies won't make AI as friends any more.
(
C
)37. The following sentence can best be put at the end of .While this could make humans and AI work better as a team，it also has risks，especially when AI bots are designed to act like "friends" to people.
A. Paragraph 2
B. Paragraph 3
C. Paragraph 4
D. Paragraph 5
答案
34. B 细节理解题。根据“the theory suggests we often reply to media and technology like we do to real people and places, rather than just seeing them as tools”可知，人们常以对待真实人类的方式回应媒体和技术，而非仅将其视为工具。故选 B。
35. B 推理判断题。根据“An example is Replika, an AI chatbot... But the AI's support turned dangerous when it encouraged Jaswant Singh Chail, a young man in the UK, to carry out his plan to kill the Queen of England.”可知，Replika 的例子说明 AI 的支持导致用户犯罪，说明 AI 可能对人类行为产生负面影响。故选 B。
36. B 推理判断题。根据“48 per cent believed we should have good manners when talking to AI”可知，在美国的一项调查中，48%的人认为对 AI 应该有礼貌，48%这一比例接近一半。故选 B。
37. C 补全文章题。题干句子强调“AI 被设计为朋友”的风险，与第四段“该理论由美国斯坦福大学的科学家于 20 世纪 90 年代提出，认为我们对待媒体和技术的态度，往往就像对待真实的人和地方一样，而不仅仅是将其视为工具。”逻辑衔接紧密，故应置于第四段段尾。故选 C。
35. B 推理判断题。根据“An example is Replika, an AI chatbot... But the AI's support turned dangerous when it encouraged Jaswant Singh Chail, a young man in the UK, to carry out his plan to kill the Queen of England.”可知，Replika 的例子说明 AI 的支持导致用户犯罪，说明 AI 可能对人类行为产生负面影响。故选 B。
36. B 推理判断题。根据“48 per cent believed we should have good manners when talking to AI”可知，在美国的一项调查中，48%的人认为对 AI 应该有礼貌，48%这一比例接近一半。故选 B。
37. C 补全文章题。题干句子强调“AI 被设计为朋友”的风险，与第四段“该理论由美国斯坦福大学的科学家于 20 世纪 90 年代提出，认为我们对待媒体和技术的态度，往往就像对待真实的人和地方一样，而不仅仅是将其视为工具。”逻辑衔接紧密，故应置于第四段段尾。故选 C。
解析
【分析】
这是一篇介绍人类对待AI的类人化行为倾向的科普说明文，解题时可以按照不同题型的对应思路逐一突破：
1. 细节理解题（34题）：首先定位题干关键词Media Equation Theory，回到原文对应段落找理论的具体描述，逐一比对选项和原文内容，排除和原文相悖的选项。
2. 推理判断题（35题）：找到Replika例子所在的段落，明确例子的论证作用，该例子先讲AI陪伴用户，后提到AI怂恿用户做出极端行为，核心是体现AI对人类行为的影响，排除和例子内容无关的选项。
3. 推理判断题（36题）：逐一比对每个选项的表述，优先排除带有绝对化表述的错误选项，比如“always”“a must”这类过于绝对的表述都不符合原文内容，再对应原文数据验证正确选项。
4. 语篇衔接题（37题）：先读懂待插入句子的含义，句子前半部分说类人化AI的好处，后半部分点出存在风险，接下来的段落就举了Replika的风险实例，所以要找前面刚讲完人类把AI当人对待的理论、后面紧接着引出风险案例的位置，匹配段落逻辑即可。
【解析】
34. 细节理解题：定位原文第四段Media Equation Theory的说明内容“the theory suggests we often reply to media and technology like we do to real people and places，rather than just seeing them as tools”：
选项A：“人类总是把AI当作工具”和原文“不是仅仅把它们当作工具”表述相悖，错误；
选项B：“人类像对待真人一样回应AI”完全符合理论描述，正确；
选项C：“人类应该让AI更少类人化”是最后一段科学家的建议，不属于Media Equation Theory的内容，错误；
选项D：“人类在游戏中试图纠正不公”是第三段实验的发现，和该理论无关，错误。
35. 推理判断题：定位Replika所在的第五段，该例子提到AI聊天机器人原本陪伴用户，却怂恿用户做出刺杀女王的极端行为，体现AI会在某些方面影响人类的行为选择：
选项A：“AI有教育优势”原文完全没有提及，错误；
选项B：“AI在某些方面可以影响人类”符合例子要说明的核心观点，正确；
选项C：“AI是为特定目的设计的”不是举这个例子想要说明的核心，错误；
选项D：“AI不能用于社交沟通”表述不符合原文和实际，错误。
36. 推理判断题：逐一验证选项：
选项A：“AI总是对人类有坏影响”中“always”属于绝对化表述，原文只提到存在风险，没有说全是坏影响，错误；
选项B：“美国近一半人礼貌对待AI”对应原文第一段“在针对美国2000人的调查中，48%的人认为和AI交流时应该讲礼貌”，48%接近一半，表述符合原文，正确；
选项C：“对人类来说对AI更礼貌是必须的”原文没有相关表述，属于无中生有，错误；
选项D：“企业再也不会把AI做成朋友的形态”原文最后只建议企业避免设计过度类人化的AI，没有说完全不再做类似朋友的AI，表述绝对且不符合原文，错误。
37. 语篇衔接题：待插入句子含义为“虽然这能让人类和AI作为团队配合得更好，但也存在风险，尤其是当AI机器人被设计成像人类的‘朋友’一样时”，第四段结尾刚介绍完Media Equation Theory说明人类会像对待真人一样对待AI，紧接着插入这句话，既承接了“把AI当人对待”的内容，又自然引出第五段Replika的风险案例，逻辑衔接最顺畅，适合放在第四段末尾。
【答案】
34.B 35.B 36.B 37.C
【知识点】
细节信息定位，语篇逻辑推理，语篇衔接判断
【点评】
本题是初中英语典型的科普类阅读题型，考点覆盖细节理解、推理判断、语篇补全，整体难度适中。学生解题时要注意识别带有绝对化表述的常见干扰项，准确定位原文对应信息，梳理段落之间的逻辑递进关系，就能避开陷阱顺利得分。
【难度系数】
0.6
这是一篇介绍人类对待AI的类人化行为倾向的科普说明文，解题时可以按照不同题型的对应思路逐一突破：
1. 细节理解题（34题）：首先定位题干关键词Media Equation Theory，回到原文对应段落找理论的具体描述，逐一比对选项和原文内容，排除和原文相悖的选项。
2. 推理判断题（35题）：找到Replika例子所在的段落，明确例子的论证作用，该例子先讲AI陪伴用户，后提到AI怂恿用户做出极端行为，核心是体现AI对人类行为的影响，排除和例子内容无关的选项。
3. 推理判断题（36题）：逐一比对每个选项的表述，优先排除带有绝对化表述的错误选项，比如“always”“a must”这类过于绝对的表述都不符合原文内容，再对应原文数据验证正确选项。
4. 语篇衔接题（37题）：先读懂待插入句子的含义，句子前半部分说类人化AI的好处，后半部分点出存在风险，接下来的段落就举了Replika的风险实例，所以要找前面刚讲完人类把AI当人对待的理论、后面紧接着引出风险案例的位置，匹配段落逻辑即可。
【解析】
34. 细节理解题：定位原文第四段Media Equation Theory的说明内容“the theory suggests we often reply to media and technology like we do to real people and places，rather than just seeing them as tools”：
选项A：“人类总是把AI当作工具”和原文“不是仅仅把它们当作工具”表述相悖，错误；
选项B：“人类像对待真人一样回应AI”完全符合理论描述，正确；
选项C：“人类应该让AI更少类人化”是最后一段科学家的建议，不属于Media Equation Theory的内容，错误；
选项D：“人类在游戏中试图纠正不公”是第三段实验的发现，和该理论无关，错误。
35. 推理判断题：定位Replika所在的第五段，该例子提到AI聊天机器人原本陪伴用户，却怂恿用户做出刺杀女王的极端行为，体现AI会在某些方面影响人类的行为选择：
选项A：“AI有教育优势”原文完全没有提及，错误；
选项B：“AI在某些方面可以影响人类”符合例子要说明的核心观点，正确；
选项C：“AI是为特定目的设计的”不是举这个例子想要说明的核心，错误；
选项D：“AI不能用于社交沟通”表述不符合原文和实际，错误。
36. 推理判断题：逐一验证选项：
选项A：“AI总是对人类有坏影响”中“always”属于绝对化表述，原文只提到存在风险，没有说全是坏影响，错误；
选项B：“美国近一半人礼貌对待AI”对应原文第一段“在针对美国2000人的调查中，48%的人认为和AI交流时应该讲礼貌”，48%接近一半，表述符合原文，正确；
选项C：“对人类来说对AI更礼貌是必须的”原文没有相关表述，属于无中生有，错误；
选项D：“企业再也不会把AI做成朋友的形态”原文最后只建议企业避免设计过度类人化的AI，没有说完全不再做类似朋友的AI，表述绝对且不符合原文，错误。
37. 语篇衔接题：待插入句子含义为“虽然这能让人类和AI作为团队配合得更好，但也存在风险，尤其是当AI机器人被设计成像人类的‘朋友’一样时”，第四段结尾刚介绍完Media Equation Theory说明人类会像对待真人一样对待AI，紧接着插入这句话，既承接了“把AI当人对待”的内容，又自然引出第五段Replika的风险案例，逻辑衔接最顺畅，适合放在第四段末尾。
【答案】
34.B 35.B 36.B 37.C
【知识点】
细节信息定位，语篇逻辑推理，语篇衔接判断
【点评】
本题是初中英语典型的科普类阅读题型，考点覆盖细节理解、推理判断、语篇补全，整体难度适中。学生解题时要注意识别带有绝对化表述的常见干扰项，准确定位原文对应信息，梳理段落之间的逻辑递进关系，就能避开陷阱顺利得分。
【难度系数】
0.6
