What is Generative AI?

Generative AI refers to a combination of technologies that use patterns from vast datasets to create new content, including written text, images, music, and videos. As a type of machine learning, these tools can assist with tasks like brainstorming ideas, automating workflows, and generating realistic visuals from simple prompts.

What sets generative AI apart from its precursors (e.g., predictive AI) is that it creatively applies patterns in new and unique ways with an uncanny ability to mimic human creativity. This advancement is largely due to increased computing power, new technological applications, and a drastic increase in the size of the datasets and parameters used for training.

Depending on the company training the model, the expansive datasets used for training generative AI models can include content from public websites, emails, chatbots, and social media sites, like Reddit, YouTube, etc.

You'll find generative AI tools embedded in many of the applications you use on a daily basis. For example, Google Search, Canva, Adobe Express, library databases, etc. These generative AI tools are usually built on generative AI models called foundation models. Some of the most popular models include GPT by OpenAI, Llama by Meta, Claude by Anthropic, Firefly by Adobe, and Midjourney.

Since different foundation models can be fine-tuned to improve performance and quality for specific tasks or subjects, generative AI tools carefully choose which model is most appropriate and may offer you the opportunity to choose which model or version of a model you would like to use.

New generative AI tools using these models crop up all the time. Of course, access to generative AI tools usually comes with a fee, often on a credit system (i.e., by number of prompts or interactions), depending on the model and version.

As a student, you'll need to consider the following risks when you choose to use generative AI tools:

• Academic integrity - Is the use of AI permitted by your professor for your assignment? Students using generative AI without authorization may be accused of plagiarism.
• Credibility - Does the response from the tool contain bias, inaccuracies, or errors? Generative AI tools use common patterns in language and images to generate content, not necessarily facts or truth. They are known to communicate errors, bias, and misconceptions.
• Copyright and intellectual property - What data has the model been trained on? Would uploading files to the tool for analysis break copyright or licensing agreements? A number of generative AI companies are currently in court for potentially illegally using copyrighted or licensed material in their dataset.
• Data security and privacy - Is the generative AI tool reputable, with high quality data security and privacy policies? Some companies creating generative AI tools may have malicious motives.
• Equality of access - Do your classmates have equal access to this tool? Many generative AI tools are fee-based or lack accessible features, creating barriers for some classmates and potentially offering you an unfair advantage.
• Environmental and human labour impacts - Do generative AI tools align with your values? The computing power required to train and develop generative AI models will have a significant impact on the environment due to high energy consumption, carbon emissions, and data centre construction. In addition, generative AI models impact human labour in a variety of ways and have the potential to disproportionately negatively affect racialized, marginalized, and disadvantaged individuals.

Remember to always verify your professor's expectations regarding generative AI tools, never share personal information in a chatbot, and always critically evaluate content produced by a generative AI tool.

Generative AI is most effective when used as a tool for enhancing and streamlining the creative process. These tools can help speed up workflows, unlock new creative ideas and assist with problem-solving.

However, AI lacks human intuition, critical thinking and originality. The best results come from combining AI generated outputs with human expertise, refinement and ethical judgement. Used thoughtfully, generative AI can be a powerful partner for creativity and research.