Become an AI Bully
TECHNOLOGY & SCIENCE
4/4/2026
WARM-UP
Answer quickly. No right or wrong answers.
Have you ever felt frustrated with technology?
Do you use AI chatbots often?
Do you think AI is reliable today?
VOCABULARY
Essential (A1–B1)
argue – to share opposite opinions, often angrily
Example: I sometimes argue with chatbots when they give wrong answers.
frustrating – making you feel annoyed or impatient
Example: It is very frustrating when technology doesn’t work.
bully – a person who scares or pushes others
Example: The job is to act like an bully and challenge the AI.
confused – not able to understand clearly
Example: The chatbot became confused after many questions.
Developed (B2–C2)
brutally – in a very direct and honest way
Example: You must speak brutally and not hide your opinions.
let down – to disappoint someone
Example: Many users feel let down by AI responses.
READING
Introduction
“This article explores an unusual job where people test AI systems by challenging them repeatedly. Focus on how and why this job exists.”
Article (A1–A2)
You ask a chatbot a question, but the answer is wrong.
You ask again, but it is still wrong.
Soon, you start to argue with it.
This can be very frustrating.
One company will pay people to do this job.
Developed Article (B1–C2)
A company called Memvid is offering $800 for people to act like an bully toward AI systems for one day. The goal is to be brutally honest about how frustrating the experience can be.
Applicants should have experience being let down by technology and must be patient enough to repeat conversations. By pushing AI systems until they become confused, the company can identify weaknesses.
Memvid focuses on improving AI memory, as many systems forget earlier parts of conversations. Studies show that AI becomes significantly less accurate in longer interactions, which explains why users often feel frustrating experiences.
COMPREHENSION CHECK
A1–A2
What happens when the chatbot gives wrong answers?
How do people feel when AI doesn’t work?
Does the company pay people for this job?
B1–B2
What kind of experience does the company want?
Why do workers repeat the same questions?
What happens when AI becomes confused?
C1–C2
Why is memory important for AI systems?
How does testing AI help improve it?
What problem do long conversations create for AI?
SPEAKING
A1–A2 — Foundation
Use: “I think… because…”
Do you get frustrated with technology? Why?
Have you ever argued with a chatbot?
Do you think AI is helpful?
What makes technology frustrating?
Do you trust AI answers?
Have you felt confused using technology?
Do you think AI makes mistakes often?
Would you do this job?
Do you like testing technology?
Is AI better than humans at some tasks?
Do you enjoy using new technology?
What is the most frustrating app you use?
Do you think AI will improve?
Would you be patient with AI errors?
Does technology make life easier?
B1–B2 — Expansion
Use: “I feel… because… for example…”
What are your thoughts on this job?
Would you be good at challenging AI?
Why do people argue with chatbots?
What makes AI frustrating?
How could AI improve?
Do you think testing AI is important?
What problems do chatbots have?
How does AI affect your daily life?
Would you trust AI advice?
How can companies improve AI systems?
Do you think AI replaces human jobs?
What are the risks of AI mistakes?
How does repetition help test AI?
Is patience important when using AI?
Do benefits of AI outweigh its problems?
C1–C2 — Depth & Reflection
Use: “I believe… because… this leads to…”
What does this job reveal about current AI limitations?
Is “bullying” AI an effective testing method?
How important is memory in artificial intelligence?
Should companies invest more in AI reliability?
How might AI errors affect critical industries?
Could repeated testing significantly improve AI systems?
What ethical concerns exist when developing AI?
How should companies handle user frustration?
Does AI dependency create new risks for society?
How could AI reshape communication in the future?
Should AI systems be transparent about limitations?
How might AI influence human critical thinking?
What role does user feedback play in AI development?
Could AI ever become fully reliable?
In your opinion, is AI currently overestimated or underestimated?
