How does the trained AI picture sorter described in the lesson make a prediction?

• It copies the label from the last photo.
• It uses mathematical rules and data to predict a label.
• It copies a label from the photo's file name.

Which example shows a mistaken AI prediction?

• A music app suggests a song the child likes after using their listening history.
• A photo sorter labels a cat photo as 'cat'.
• A voice helper hears 'call Tom' after the child says 'call Mom'.

Missing examples or mistaken labels in training data can lead to AI mistakes.

A dog sorter saw many retrievers but very few poodles during training. It mistakes a poodle photo for another animal. Which training problem could explain this?

• It had few poodle examples.
• It learned mainly from poodles rather than retrievers.
• It trained on dog names instead of the dog photos.

Zara typed 'I will meet you soon'. Her phone's autocorrect changed 'soon' to 'moon'. What happened?

• A voice helper misheard a spoken word.
• Autocorrect changed a typed word to an unintended word.
• A photo sorter gave a photo an unintended label.

A recommendation tool uses viewing history. After a child watches a maths video, it suggests a maths app. Which clue is stated in the question?

• The maths app's position on the shop's front page.
• The child's typed search for a maths app.
• The child's earlier viewing of a maths video.

Several apple photos in a training set were mistakenly labeled 'pear'. Why could this cause later mistakes?

• They can add pear photos to the training pictures.
• They can teach an apple-to-pear connection.
• They can make fruit counting replace the sorting task.

A sorter sometimes confuses muffins and puppies. You do not yet know why. What should you do to investigate?

• Test new photos and compare guesses with known labels.
• Add more copies of the original training photos.
• Change the known photo labels to agree with the sorter's guesses.