The Basics of Generative Artificial Intelligence
The University of British Columbia has posted the following definition of Generative AI:
"Generative AI refers to artificial intelligence systems that can generate new content—such as texts, images, audio, and video—in response to prompts by a user, after being trained on an earlier set of data."
It's very important to understand how Generative AI platforms, like ChatGPT, differ from previous digital information search engines such as Google. The key difference is that 'Googling' something involves quickly locating a website (or other digitally available information) that already exists in the form in which you view it. Something found through Google by one user can also be found, exactly in the same form, by another user. Conversely, a prompt on ChatGPT will result in new and unique content, based on a massive database, and what is generated or 'found' is unlikely to be generated in exactly the same way for someone else.
As of early 2024, arguably the most common AI tool is ChatGPT. Numerous other examples are becoming popular, and they include DALL-E, TOME, Synthesia, Google Bard, Bing Chat and more. Recently a number of education-specific AI tools have arrived, one popular example being 'Magic School'. Some Generative AI platforms require a paid subscription, while others do not. It's important to note that we are at the very beginning phase of this technology - we can expect this to evolve.
It could be argued that Generative AI is a more like a ‘system’ than a device. This is not VHS or Beta, or Xbox versus Play Station. Artificial Intelligence is not a brand, though there are many companies that use it. According to IBM, AI is a system or field, 'which combines computer science and robust datasets to enable problem-solving'.
Generative AI can be incredibly powerful, but it does have clear limitations. Here are some strengths and weaknesses:
Strengths of AI:
Platforms like ChatGPT can do a myriad of things such as:
- Pass exams like SATs and medical tests
- Solve Math problems and break them down by steps
- Summarize, analyze and re-write texts
- Translate languages and write letters in virtually any foreign language
- Compose multiple versions of a prompt or idea
Limitations of AI:
Platforms like Chat GPT may not be able to:
- Provide accurate or correct responses
- Predict future events
- Cite sources (it might make them up!)
- Relate to personal events or experiences
- Connect to community customs and local/oral history
Considering the strengths and limitations above, as well as the fact that this technology is very new and it's social and community impacts still remain to be seen, school districts may want to take a position similar to that of the University of British Columbia. According to the UBC page on Generative AI, they state:
'UBC is taking a measured and balanced approach to the use of generative AI tools in teaching and learning as these tools present both opportunities and risks.'