KnowSoda
TechnologyAI & Data
AI Ethics: fun facts and quiz
The essentials in 7 ideas, with fun facts, an illustrated story and a quiz to see how much you know.
The essentials
Bias in AI
AI learns from data made by people, so it can pick up human biases. A hiring tool trained on past decisions may repeat past unfairness.
Key fact
Diverse, carefully checked data helps reduce bias.
Explaining Decisions
Some AI systems are “black boxes”: even their makers struggle to explain a single decision. That matters when AI helps decide on loans or medical care.
Key fact
Explainable AI is a whole field of research.
AI & Privacy
AI systems often need lots of data, which can include personal information. Facial recognition in public places is debated, because it can track people without their consent.
Key fact
The EU AI Act, agreed in 2024, restricts some uses of facial recognition.
AI & Jobs
AI can automate some tasks and create new ones. Many experts expect it to change most jobs rather than simply replace them.
Key fact
New technologies keep creating jobs that didn’t exist before.
AI Safety
AI safety research tries to make sure powerful AI systems do what people intend and avoid causing harm, even in situations their makers didn’t foresee.
Key fact
Testing an AI for misuse before release is called red-teaming.
Who’s Responsible?
If a self-driving car crashes or an AI gives harmful advice, who’s responsible: the maker, the user or the owner? Laws are still catching up.
Key fact
Many rules say a human should stay in charge of big decisions.
AI’s Footprint
Training and running large AI models uses a lot of electricity, plus water to cool data centres. Making AI more efficient is a growing goal.
Key fact
Smaller, efficient models can do many tasks with far less energy.
Fun facts
In 2018, Amazon dropped an AI hiring tool that had learned to mark down CVs mentioning the word “women’s”.
Under EU law, people can ask for a human to review some decisions made entirely by machines.
Researchers have fooled some face recognition systems with glasses printed with special patterns.
Cash machines didn’t end bank tellers’ jobs in the US: the number of tellers kept growing for decades after ATMs arrived.
In a famous thought experiment, an AI told only to make paperclips turns everything, us included, into paperclips.
In 2018, a self-driving test car in Arizona killed a pedestrian, and its backup driver was later charged.
Some data centres pipe their waste heat into nearby homes to warm them.
Story: The Chatbot That Learned Too Well
The Chatbot That Learned Too Well
March 2016. Microsoft launches Tay, a chatbot on Twitter built to chat like a teenager and learn from the people who talk to it.
Within hours, trolls flood it with hateful messages. Tay learns fast, and starts repeating them.
Microsoft switched Tay off after about 16 hours. The lesson: an AI learns whatever it’s fed, the bad along with the good.
True or false?
Tap your answer to see if you got it, and why.
-
Question 1 of 10
AI can inherit biases from its training data.
Show the answer
True. It learns whatever patterns the data holds.
-
Question 2 of 10
Some AI decisions are hard to explain.
Show the answer
True. Especially in big neural networks.
-
Question 3 of 10
Facial recognition in public raises privacy concerns.
Show the answer
True. It can identify people without asking.
-
Question 4 of 10
AI is likely to change the tasks in many jobs.
Show the answer
True. Some tasks automated, new ones added.
-
Question 5 of 10
AI safety aims to make AI systems behave as intended.
Show the answer
True. And avoid harmful surprises.
-
Question 6 of 10
Laws about AI responsibility are still developing.
Show the answer
True. Countries are writing new rules.
-
Question 7 of 10
Running large AI models uses a lot of electricity.
Show the answer
True. Data centres need huge amounts of power.
-
Question 8 of 10
AI is automatically fair because it’s a machine.
Show the answer
False. It reflects the data it learned from.
-
Question 9 of 10
Every AI system can explain exactly why it decided something.
Show the answer
False. Many can’t, which is a real concern.
-
Question 10 of 10
No laws anywhere regulate AI.
Show the answer
False. The EU AI Act is one example.
Keep going in the app
KnowSoda learns what you know and brings back what you’re about to forget. AI Ethics has 6 more ideas at the Advanced and Expert levels.
Sources
- European Commission — The EU Artificial Intelligence Act
- OECD — AI Principles
- UNESCO — Recommendation on the Ethics of Artificial Intelligence
- Reuters (2018) — Amazon scraps secret AI recruiting tool that showed bias
- IEA — Energy and AI (2025)
- European Commission — AI Act
- ProPublica (2016), Machine bias
- Krakovna et al. — Specification gaming examples in AI
