AI (artifical intelligence) is here – well, everywhere actually.

With the launch of ChatGPT just under a year ago, AI made its presence more known in classrooms. ChatGPT is not the only AI in town, but it currently it is one of the better known generative AI applications. 

ChatGPT, a project of OpenAI,  is financed by Microsoft. Google Bard is a competitor – financed by, you guessed it, Google.  

New applications – including many based on OpenAPI – are being released daily.

What is generative AI anyway?

Generative AI is not taking information “in its entirety” and giving it to you the way Siri or Alexa would.  It is generating, or creating, the content it gives you. This content can include text but it can also include visual works such as images.

Generative AI is based on a large language model (LLM) – meaning it is based on large amounts of text data, mostly taken from the Internet in mid-2020. This text has been scrubbed – in other words, cleaned up to remove irrelevant and inappropriate content. The LLM has also been tested and trained to provide accurate, coherent, and contextually appropriate responses.

ChatGPT as well as Google Bard and other generative AI applications are chatbots – meaning you ask questions and it answers. In other words, it is conversational.

What are important things to consider with generative AI?

With chatbots, it is important how and what you ask in your question – or prompt. The art of writing good prompts, prompts that give you accurate, coherent, and contextually appropriate responses is known as prompt engineering. Prompt engineering is a skill that must be learned and taught to our students.

Even with well-trained LLMs and good prompts, generative AI can create content that is wrong.  These are referred to as hallucinations. Unless the persons creating content with AI has an understanding of the content, these hallucinations can go undetected – or worse, be considered to be “the truth”.

With the release of ChatGPT, red flags immediately went up in the education world about the use of generative AI to complete classroom assignments. Cheating, is, of course, possible.  However, how well a student can use generative AI includes how well they can write prompts and how well they review the content that is created. It is also impacted by the specific assignment given. A number of products have been released that claim they can detect whether or not text was written by AI. But they don’t always work. They can fail to recognize AI output, especially if it is edited, and they can misidentify human-written text as AI-generated.

How can generative AI help you?

Generative AI can be used to save you as an educator time, particularly for grading  and creating instructional materials.

A good source of information about this is “Using AI Chatbots to Enhance Planning and Instruction (Quick Reference Guide)” by Monica Burns.

What do you do next?

Learn more

• I found “The AI Classroom: The Ultimate Guide to Artificial Intelligence in Education (The Hitchhiker’s Guide for Educators Series)” by Dan Fitzpatrick, Amanda Fox, and Brad Weinstein to be a great source of information
• The challenge here can be to learn more without being consumed by the vast amount of information out there.

Test it out yourself

• Get an account – there are free accounts available for both ChatGPT (openai.com/chatgpt) and Google Bard (bard.google.com)
• Ask questions, even (and maybe especially) ones you know the answer to

Set goals and expectations

• Set goals for how you will learn about and use generative AI
• Set expectations for how your students can – and can not – use generative AI

Prepare for all this to change – fast!

• Just last month, it was announced that Google Bard can interact with real time flight and hotel information and be enabled to interact with information from your gmail, google docs and google drive.
• What’s next?