AI 논쟁 수업 - 1st Article

2023년, 타임지에서 오픈AI 발전을 위해 케냐의 노동자들이 어떤 고초를 겪고 있는지에 대한 기사를 접했을 때, 기술 발전에서 이런 이슈를 모른척 하지 않도록 수업에서 다루어보고 싶었다.  작년에는 진로영어를 하느라 다룰 기회가 없었고, 올해는 새 학교에서 동료와 협의도 쉽지 않고, 주당 2시간인 내 수업에서 교과서 외 지문을 다룰 시간적 여유가 부족해서 엄두를 못 내고 있던 차에, 국어과 선생님과 IB 교학공에서 수업 계획을 나누다가 죽이 맞아 융합수업을 설계하기로 했다. 논제를 정하고 자료를 찾으며 주장을 다듬는데 시간이 많이 걸리는데 그 과정을 두 과목에서 함께 진행하면 서로의 시간을 줄여주면서 두 언어로 사고하고 표현하는 과정에서 더 풍부한 배움이 일어나지 않을까 기대한다. 

 

국어 시간에 [주장하는 글쓰기] 수행평가 과업을 설정하였고, 구본권, 《로봇 시대, 인간의 일》 을 함께 읽고 인공지능에 대한 생각, 경험, 배경지식, 궁금증을 나눈다. 더 알아보고 싶은 탐구 주제를 찾고 질문을 만든 후 탐구할 논제를 선택하고 논제를 둘러싼 다양한 의견 조사·분석 과정을 거쳐 주장을 정하고, 근거를 마련하고, 개요를 작성하고, 주장하는 글을 쓴다. 

 

영어시간에는 인공지능의 발전과 관련된 딜레마를 다루는 기사를 3개 학습한 후 모둠별로 기사를 한국어판으로 제작하는 활동을 통해 비판적 시각으로 기사를 재구성해보고 각 모둠이 뽑은 헤드라인 및 기사 내용의 차이를 살펴보며 같은 기사 원문에서 견해가 어떻게 달라질 수 있는지 확인한다.  3개 기사 중 국어 시간과 연계한 논제를 설정할 수 있으면 더욱 좋겠다. 국어시간에 쓴 글을 바탕으로 Plang이라는 말하기 도구를 활용해서 피드백을 받아가며 자신의 주장을 말하는 연습을 한다.  주장을 강조하고, 논리를 전개하며, 청중을 설득하는 데 도움이 되는 다음 표현을 학습할 예정이다.

 

1. 주장 제시하기

• I strongly believe that...
• In my opinion...
• From my perspective...
• I am convinced that...
• It is my firm belief that...

2. 논리 전개하기

• Firstly... Secondly... Finally...
• Moreover... / Furthermore...
• In addition to this...
• Not only... but also...
• On the one hand... On the other hand...

3. 예시 제시하기

• For instance...
• To illustrate this point...
• Let me give you an example...
• Consider the case of...
• As an illustration...

4. 강조하기

• It is crucial to note that...
• I would like to emphasize that...
• The key point to remember is...
• What's particularly important is...
• We must not forget that...

5. 반대 의견 다루기

• Some might argue that... However...
• While it's true that... we must also consider...
• Admittedly... Nevertheless...
• It's understandable to think that... but...
• Despite the common belief that... in fact...

6. 결론 내리기

• In conclusion...
• To sum up...
• In light of these facts...
• Considering all these points...
• Based on the evidence presented...

7. 행동 촉구하기

• I urge you to...
• It's time for us to act...
• We must take steps to...
• Let us not wait any longer to...
• The time has come for us to...

8. 청중과 공감대 형성하기

• As many of you may know...
• Like many of you, I too...
• We've all experienced...
• I'm sure you'll agree that...
• Let's consider for a moment...

9. 신뢰성 확보하기

• According to recent studies...
• Experts in the field have found that...
• Research has shown that...
• Statistics indicate that...
• As demonstrated by...

10. 감정에 호소하기

• Imagine for a moment...
• Think about how you would feel if...
• Consider the impact on...
• We cannot ignore the plight of...
• The human cost of this issue is...

[ Sample Output ]

 

The Hidden Cost of AI Development: What We Need to Do

ChatGPT is an amazing AI. But in my opinion, there's a hidden problem behind this success. 
   First, OpenAI paid Kenyan workers less than $2 per hour for their work. For example, these workers had to read bad content like violence or abuse for 9 hours a day. This is not fair.
   Second, this job is bad for the workers' mental health. We must remember this. Of course, some people might say this job gives opportunities to Kenyan people. However, jobs shouldn't harm people's health. What I want to emphasize is that AI companies are making a lot of money while giving dangerous work to people in poor countries. This is not right.
  In conclusion, AI is important, but it should treat all people fairly. We need to act now. We should demand that AI companies treat all workers fairly.
  Everyone, please care about this issue. If we speak up, we can make a change.

 

기사 원문이 어려워 Briskteaching 의 레벨링을 했는데 작업 결과물이 썩 만족스럽지 않아 클로드와 함께 학생들에게 적절한 난이도로 기사를 수정하였다.

*원문에는 너무 끔찍한 묘사가 포함되어 있어서 행여 학생들이 읽을까봐 여기에는 출처를 밝히지 않습니다. 

 

 

** Material : 수업용으로 각색한 기사

OpenAI Used Low-Paid Kenyan Workers to Make ChatGPT Safer: A Closer Look

 

The AI Success Story

ChatGPT, created by OpenAI, became a huge hit in late 2022. This AI chatbot can write about almost anything, from poetry to explaining complex science. It got over a million users in just a week! OpenAI is now worth billions of dollars and might get a big investment from Microsoft.

 

The Hidden Story

But there's more to this success than just smart people in Silicon Valley. TIME magazine found out that OpenAI used workers in Kenya, paying them less than $2 per hour, to make ChatGPT less toxic.

 

Why Was This Work Needed?

ChatGPT's earlier version, GPT-3, was really good at putting sentences together. But it had a big problem: it would sometimes say violent, sexist, or racist things. This happened because it learned from billions of words from the internet, including the bad stuff.

 

The Solution and Its Cost

To fix this, OpenAI needed people to label toxic content. They hired a company called Sama in Kenya to do this work. Starting in November 2021, Sama workers had to read and label tens of thousands of pieces of text. Some of this text was really disturbing, describing things like child abuse, violence, and other harmful content.

 

The Workers' Experience

The Kenyan workers were paid between $1.32 and $2 per hour. They had to read between 150 and 250 text passages during a 9-hour shift. Many workers said this job affected their mental health. While they could talk to counselors, workers said these sessions weren't very helpful.

 

The Ethical Dilemma

This situation raises important questions:

1. Is it fair to pay workers so little for such difficult work?
2. Is it okay to expose people to harmful content to make AI safer?
3. Who really benefits from AI advancements?

The Bigger Picture

This story shows a hidden side of AI development. While companies like OpenAI make billions, they rely on low-paid workers in poorer countries to do tough, sometimes traumatizing work. It's a reminder that behind every AI success, there might be human costs we don't see.

 

What Happened Next?

In February 2022, Sama stopped working with OpenAI because of how upsetting the work was. This meant some workers lost their jobs or got paid less. OpenAI said they take mental health seriously, but they let Sama manage the workers.

 

Why This Matters

As AI becomes more important in our lives, it's crucial to think about how it's made. This story helps us understand that creating "safe" AI often involves real people doing difficult work. It challenges us to consider the ethics of AI development and who pays the price for technological progress.

 

 

[혹시 쓸데가 있을까 싶어 남겨놓는 기타 자료] 클로드 제작 lesson plan  [The Ethics of AI Development]

**Target Level:** 11th Grade English Class     

## Lesson Objectives:
1. Comprehend and analyze the key points of the article about OpenAI's use of Kenyan workers.
2. Develop critical thinking skills by discussing ethical implications of AI development.
3. Improve persuasive speaking skills through a debate activity.

## Lesson Outline:

### 1. Introduction (10 minutes)
- Ask students what they know about ChatGPT or AI chatbots.
- Briefly introduce the topic of AI ethics.

### 2. Reading and Comprehension (20 minutes)
- Distribute the article and have students read it individually.
- In pairs, students discuss and write down 3-5 key points from the article.
- Class discussion: Create a mind map on the board with the main ideas from the article.

### 3. Vocabulary Focus (10 minutes)
- Highlight key terms: AI, chatbot, ethics, toxic content, outsourcing.
- Students create sentences using these terms to demonstrate understanding.

### 4. Ethical Discussion (15 minutes)
- In small groups, students discuss the ethical questions raised in the article:
  1. Is it fair to pay workers so little for such difficult work?
  2. Is it okay to expose people to harmful content to make AI safer?
  3. Who really benefits from AI advancements?
- Groups share their thoughts with the class.

### 5. Persuasive Speaking Activity: AI Ethics Debate (30 minutes)
- Divide the class into four groups, assigning each group one of these positions:
  A. OpenAI's actions were justified for technological progress.
  B. OpenAI's actions were unethical and exploit vulnerable workers.
  C. The Kenyan government should regulate such work to protect workers.
  D. Tech companies should find alternative methods to improve AI safety.

- Preparation (10 minutes): Groups prepare arguments for their assigned position.
- Debate Format (20 minutes):
  - Each group presents their opening argument (2 minutes each)
  - Groups can challenge other positions (5 minutes)
  - Each group presents closing arguments (1 minute each)

### 6. Reflection and Conclusion (5 minutes)
- Students write a brief reflection on what they learned and how their perspective on AI development has changed.

## Extension Activities:
1. Essay writing: Students compose a persuasive essay on their personal stance regarding the ethical issues discussed.
2. Research project: Students investigate other ethical concerns in technology development and present their findings.

## Assessment:
- Participation in discussions and debate
- Quality of arguments in the debate
- Written reflection at the end of class