ARTIFICIAL INTELLIGENCE · LESSON 7 OF 8
What AI can get wrong
AI is useful, but even advanced systems can fail. Understanding their limits makes it easier to benefit from them safely and realistically.
Read the explanations, explore the activity and answer the knowledge check. No programming knowledge is assumed.
Confident but incorrect answers
Some generative AI systems produce false statements or references that sound plausible. This is sometimes called hallucination. Check important facts against credible sources.
Bias and privacy
Training data can contain patterns of unfairness or errors. Sending personal information to an AI service may also introduce privacy risks depending on its terms, storage and access controls.
Security and misuse
AI tools can be misused for impersonation, fraud and malicious content. Automated systems can also make harmful mistakes if given excessive permissions without appropriate safeguards.
Human oversight
Decisions affecting people, sensitive data or critical systems deserve proportionate human review, monitoring and independent testing.
Practise checking an output.
“Research proves this tool is always accurate.”
A confident statement gives us a claim to investigate. It supplies no evidence by itself.
0 of 4 review prompts considered. This checklist does not certify a system as safe or accurate. Selections stay on this page and are not sent anywhere.
COMPUTING FOUNDATIONS · STEP 11 OF 13
Check what you learned
Answer both questions correctly to complete this step. You can revise your answers and retry.
Progress stays in this browser. It is not synced to an account or sent as analytics. Passing these introductory checks is not certification.
Explore another AI topic
About this lesson
Half Echo Learn explanation and introductory activity. Educational illustrations do not represent a running AI service. The knowledge check tests the concepts on this page.