Reggie the Lion, the precious mascot of King’s College London (KCL), turns 100 this month, and previously this year he ended up being the muse for my very first venture into utilizing expert system (AI) in the class.
I chose to utilize Reggie and this year’s most talked-about innovation to develop illustrations for my lecture products. I’m based in KCL’s organization school, where I’m an assistant teacher in marketing innovation and development. Formerly I would have counted on generic stock images, however on this event, I utilized generative AI– particularly text-to-image designs that take natural language as input and produce images as output– to breathe life into bespoke illustrations, producing custom-made material for my trainees.
In my viewpoint, there’s still great deals of unpredictability and apprehension about AI in college, especially amongst instructors. Numerous are uncertain about how to harness its capacity. In a study of scientists previously this year, Nature discovered that majority believed that generative AI would make it more difficult to examine trainee knowing.
AI and science: what 1,600 researchers think
After experimenting with AI tools, it didn’t take wish for me to make Reggie the lion a crucial part of my mentor slides. When I discussed a mental research study as part of my curriculum on customer behaviour, there was Reggie, putting on an electroencephalogram headset, which determines the electrical activity of the brain. Reggie was walking through a grocery looking for the finest product when we talked about retail and product placement.
Using generative AI to develop images can develop problems in copyright law. In my case, KCL has the rights to its own mascot, and since I’m utilizing these images for internal mentor functions, I’ve yet to experience any copyright concerns. ( Nature does not enable the usage of generative AI to develop images for its pages.)
My usage of generative AI may not make the headings as significantly as trainees utilizing ChatGPT to compose coursework, however we rely greatly on slides in college and usage images to magnify the material. I argue that enhancing visuals, to make them more relatable and individual is an engaging usage case for generative AI.
What I utilize
I’ve try out OpenAI’s tool DALL-E, Stability AI’s Stable Diffusion and Midjourney text-to-image designs. When I began messing around in the location in 2022, a few of the services still needed coding and understanding of computer system facilities to get them to work; now they merely need a log-in.
All of the designs can develop photorealistic pictures of human beings, however they have a hard time more when it pertains to images that wasn’t in the algorithms’ training information– a lion using a red T-shirt, for instance, can be a huge ask. After having a hard time to produce constant images, I began to evaluate different illustration designs. I’ve had the most success with a method simulating the computer-animation design of Pixar. What works best depends upon the wanted material and context– experimentation is needed.
For the trainees and me, the ‘AI Reggie project’ wasn’t simply an innovative workout; by imagining abstract ideas through individualized AI-generated art, I triggered trainees’ interest and attention to spike. As a trainer, I discovered that the images triggered happiness and interest, and bridged the space in between private trainees’ journeys and the abundant tapestry of our organization’s history. It likewise amassed trainees’ interest in the real innovation, with lots of checking out the tech and welcoming frontier themselves.
I prepare for Reggie to end up being a more active part of trainee life. The initial images individualized the experience, made it remarkable and brought a sense of spirit and connection to the university mascot, according to a brief study of my trainees. For me, the next action is to co-create with them. I hope that this will assist to reinforce my lecture products.
This is a short article from the Nature Careers Community, a location for Nature readers to share their expert experiences and suggestions. Guest posts are encouraged.
The author states no completing interests.