Academic Integrity Tutorial: The Big Picture

We live in a world where advanced technologies known as generative artificial intelligence, or GenAI, can produce art, music and even stories that are almost indistinguishable from those created by humans. Popular examples include ChatGPT, DALL-E and MuseNet. These tools don't only replicate existing work but are designed to generate new or original content based on the patterns learned from the data they are trained on.¹

If this is true, what might the implications be of machine-generated content for artists, musicians, or journalists whose work is used to train GenAI, and for students who take advantage of the content created by GenAI? Some important considerations are

• Who owns AI-generated content?
• How accurate or reliable is AI-generated content?
• What are the ethical implications of using AI-generated content?

Now, consider another aspect - the darker side of AI-generated content. What if these same machines can be used to manipulate information, spread misinformation, or create deceiving deep fakes? As you ponder the world of AI-generated content, ask yourself

• What risks are associated with bias or misinformation in AI-generated content?
• Is Gen AI plagiarizing when it produces content that imitates existing work?
• How does the use of AI in content creation affect the job market for human content creators?
• Is there a need for regulation and oversight of AI-generated content to ensure its responsible use?

Before we dive in, perhaps the most fundamental question is, how do machines learn to generate content?

How do Machines Learn?

In this video, computer scientist Hilary Mason explains machine learning in five levels of increasing complexity by talking about it to 5 different people: a child, a teen, a college student, a grad student and an expert. It is a long video so you can choose to watch one or two interviews or the entire video if you have time. It can easily be watched at 1.5 times the speed.

Source: Computer Scientist Explains Machine Learning in 5 Levels of Difficulty by WIRED on YouTube

Test Your Understanding

Attempt this quiz even if you watched only part of the video. For each question, more than one response may be applicable but only the one mentioned in the video will be marked as correct. Click the blue button and read the feedback provided after each question; it confirms what is discussed in the video. After reviewing the feedback, click the blue arrow to 'continue'. This quiz on machine learning consists are five questions.