As part of University of the Arts London’s 12 Days of AI course, I was required to create a presentation using the AI service Gamma.
This turned out to be an fine example of using AI as a helper to overcome various work-related ailments: procrastination, writer’s block and a wasteful propensity to switch back-and-forth between writing and presentation tasks. Hailing from Adam Smith’s stomping ground, you’d think I could manage to divide up my labour effectively…….
I am currently updating guidance on the procurement of e-learning content. The current guidance is text-based and rather dull. Having identified a helpful acronym with which to arrange the content, I was in the no man’s land of having neither completed the content or settled on an effective way of presenting it. A perfect opportunity to use Gamma, perhaps?
On opening Gamma for the first time I opted for an extended prompt:
Create a presentation to provide advice on buying e-learning content. Use the following headers: standards, content, accessibility…….etc..
I was pleasantly surprised with the results. Within seconds I had a structured presentation which was more than visually acceptable, and with credible and accurate information – albeit a little generic and repetitive.
I was then able to quickly delete some slides, change headings and replace some of the content with my own. A little more tweaking to the colours and fonts and I was happy – because I had moved quickly from 80% of the content and no real idea how to present it to something that was almost complete! So, as a subject matter expert who was struggling to present my knowledge I found this activity and service highly productive. It also provided a lot of opportunity for reflection.
Generative AI is often cited as a way to get started on a piece of written work. In this case it allowed me to (almost) finish it. I was quickly able to identify the content I wanted to keep, and that I needed to discard. Blandness and a little repletion were then main issues here, rather than accuracy. The importance of existing knowledge, experience and criticality remain key to the successful use of services such as Gamma. I would have had to spend a lot more time checking the relevance, provenance and accuracy on a topic such as Christmas baubles (Bah, humbug). Users should not be bedazzled by the bells-and-whistles and thereby forget that this is still generative AI with a shiny front-end.
Templates can lead to homogenous content, and if I was marking 200 presentations made in this way I would soon become fatigued. A counter argument to that might be that a certain uniformity affords closer analysis of the content. When the content is being assessed rather than presentation this may work well for students and markers.
Having downloaded the presentation to PowerPoint and PDF I found that there were repeated accessibility issues. In particular, alt-text was missing from the photographs and various shapes used to present the content. I was disappointed by this gap in provision, given that some AI services are reasonably adept at creating alt-text. Perhaps I should have included it in my prompt. Writing alt-text is very much context driven activity but a little assistance here would have helped to kick start it.
So, my takeaway is use this not to save time but to use time more effectively.
