Why read this: This short text introduces a big idea from today's news. AI companies say their tools are dangerous, but they keep selling them. The reading helps students notice when words and actions do not match.
What to notice: Look for the main claim in the first paragraph. Then find one expert's view in the fourth paragraph. Notice which sentences are facts about companies and which sentences give an opinion about why they act this way.
Skills practised: Students practise finding the main idea in a short article. They practise picking out supporting evidence, like names and quotes. They also practise telling a company's view apart from a critic's view.
AI companies say their tools are scary, but they still sell them
Why do AI makers warn us about their own products?
Tap any green word in the article to see its meaning.
AI companies build very powerfulpowerful: Having a lot of strength, influence, or control. — 强大的；有力的 AI toolstool: An object used for doing a particular job or activity, such as a hammer or knife. — 工具；器具. They say these tools are so strong that they could harm the world. They warnwarn: to tell someone about a possible danger or problem — 警告；提醒 us that AI could eveneven: Exactly, just, fully — 正好 destroydestroy: To damage something so badly that it cannot be used or no longer exists. — 破坏；摧毁 people one day. But here is the strangestrange: Unusual or surprising in a way that is difficult to understand. — 奇怪的；陌生的 part. The same companies keep selling these tools to everyone, every day.
One AI company is called Anthropic. It made a new AI tool named Claude Mythos. Anthropic says Mythos can find weakweak: Not having much physical strength or power. — 虚弱的；软弱的 spotsspot: a particular place or area; a small round mark — 地点；斑点 in computers better than toptop: Best; of the highest quality or rank — 顶尖的；最好的 expertsexpert: A person who knows a lot about a subject — 专家. The company warns that this powerpower: The ability to control things, or energy like electricity. — 力量；权力；电力 could be very dangerous in the wrong hands: held by people who would use it to harm others — 落入恶人手中. Even so, Anthropic still uses the tool and shares it with partners.
Another big AI company is OpenAI. Its leaderleader: A person who is in charge of a group. — 领导者；领袖, Sam Altman, has often spoken about danger too. In 2015, he said AI could leadlead: to cause or encourage someone to do something; to result in a particular outcome — 导致 to the end of the world. Later, he said he loses sleep at night because he worriesworry: To feel anxious or troubled about a problem or possible event. — 担心；烦恼 about ChatGPT. Still, OpenAI keeps making bigger and stronger AI tools each year.
Some experts think this is a cleverclever: Quick at learning and understanding things; intelligent. — 聪明的；机灵的 tricktrick: Something designed to fool or swindle — 诡计;骗局. Shannon Vallor is a professorprofessor: A senior teacher at a university. — 教授 in Scotland who studies AI. She says these companies want people to feel small and afraid. When we feel afraid, we believe only the AI companies can keep us safesafe: Free from risk — 无风险的. Some experts call this fear mongering: spreading fear on purpose to gain something — 贩卖恐惧. This kind of talk can pushpush: To move something away from you using force. — 推；推动 the price of the company up and stop new safetysafety: The state of being safe and protected from danger or harm. — 安全；安全性 rules.
So what should you think about all this? Listen to what AI companies say about danger. Then ask one simplesimple: Uncomplicated; taken by itself, with nothing added — 简单的；不复杂的 question. If a tool is so dangerous, why do they keep selling it? Their loudloud: Making a lot of noise. — 大声的；嘈杂的 warningswarning: a statement or event that tells you something bad may happen — 警告；警示 and their quiet salessale: The activity of selling something, or a time when prices are lower. — 出售；销售；打折 tell two very different stories.
AI companies build very powerfulpowerful: Having a lot of strength, influence, or control. — 强大的；有力的 AI toolstool: An object used for doing a particular job or activity, such as a hammer or knife. — 工具；器具. They say these tools are so strong that they could harm the world. They warnwarn: to tell someone about a possible danger or problem — 警告；提醒 us that AI could eveneven: Exactly, just, fully — 正好 destroydestroy: To damage something so badly that it cannot be used or no longer exists. — 破坏；摧毁 people one day. But here is the strangestrange: Unusual or surprising in a way that is difficult to understand. — 奇怪的；陌生的 part. The same companies keep selling these tools to everyone, every day.
One AI company is called Anthropic. It made a new AI tool named Claude Mythos. Anthropic says Mythos can find weakweak: Not having much physical strength or power. — 虚弱的；软弱的 spotsspot: a particular place or area; a small round mark — 地点；斑点 in computers better than toptop: Best; of the highest quality or rank — 顶尖的；最好的 expertsexpert: A person who knows a lot about a subject — 专家. The company warns that this powerpower: The ability to control things, or energy like electricity. — 力量；权力；电力 could be very dangerous in the wrong hands: held by people who would use it to harm others — 落入恶人手中. Even so, Anthropic still uses the tool and shares it with partners.
Another big AI company is OpenAI. Its leaderleader: A person who is in charge of a group. — 领导者；领袖, Sam Altman, has often spoken about danger too. In 2015, he said AI could leadlead: to cause or encourage someone to do something; to result in a particular outcome — 导致 to the end of the world. Later, he said he loses sleep at night because he worriesworry: To feel anxious or troubled about a problem or possible event. — 担心；烦恼 about ChatGPT. Still, OpenAI keeps making bigger and stronger AI tools each year.
Some experts think this is a cleverclever: Quick at learning and understanding things; intelligent. — 聪明的；机灵的 tricktrick: Something designed to fool or swindle — 诡计;骗局. Shannon Vallor is a professorprofessor: A senior teacher at a university. — 教授 in Scotland who studies AI. She says these companies want people to feel small and afraid. When we feel afraid, we believe only the AI companies can keep us safesafe: Free from risk — 无风险的. Some experts call this fear mongering: spreading fear on purpose to gain something — 贩卖恐惧. This kind of talk can pushpush: To move something away from you using force. — 推；推动 the price of the company up and stop new safetysafety: The state of being safe and protected from danger or harm. — 安全；安全性 rules.
So what should you think about all this? Listen to what AI companies say about danger. Then ask one simplesimple: Uncomplicated; taken by itself, with nothing added — 简单的；不复杂的 question. If a tool is so dangerous, why do they keep selling it? Their loudloud: Making a lot of noise. — 大声的；嘈杂的 warningswarning: a statement or event that tells you something bad may happen — 警告；警示 and their quiet salessale: The activity of selling something, or a time when prices are lower. — 出售；销售；打折 tell two very different stories.
Questions
Check your understanding
- 01
What is the strange part the article points out about AI companies?
- 02
What does Anthropic say about its new tool, Claude Mythos?
- 03
Why does Shannon Vallor say AI companies talk about danger?
- 04
Describe the main claim of this article in your own words.
Suggested length: ~50 words
- 05
List three things AI companies do that the article finds strange or worrying.
Suggested length: ~50 words
Questions
Check your understanding
- 01
What is the strange part the article points out about AI companies?
- 02
What does Anthropic say about its new tool, Claude Mythos?
- 03
Why does Shannon Vallor say AI companies talk about danger?
- 04
Describe the main claim of this article in your own words.
Suggested length: ~50 words
- 05
List three things AI companies do that the article finds strange or worrying.
Suggested length: ~50 words