My students need to be well versed in AI in order to be industry ready, fashion lecturer says


A fashion lecturer is pushing the use of Artificial Intelligence (AI) in his teaching to ensure students at De Montfort University Leicester (DMU) have the cutting-edge skills needed to succeed in a multi-billion-pound industry.

ALAN BEATTIE - main pic

Alan in front of some of his AI-assisted fashion pictures

Alan Beattie, Programme Leader of Fashion Textile Design at DMU, says his students need to embrace AI as it gets bigger and faster so they are prepared for an industry that ‘has always moved incredibly quickly’.

He said: “I do not think our students will be industry ready unless they are well versed in AI technology.”

Alan says people should look at AI in fashion as an extra tool and use it for everyday tasks such as creating sketches and images.

“We need to break down the barriers of resistance to AI,” he says. “It is just a tool. It is no different to what we teach with anyway. A pencil is a tool, a paintbrush is a tool, a pair of scissors is a tool.

“We are still creating, working creatively and using our knowledge of design culture to get a resulting image from AI.”

AI relies on people writing ‘prompts’ which it then uses to provide an image incorporating all of those words.

ALAN - icarus resize

One of the photos created by Alan. Some of the prompts included The Flight of Icarus; The concept of escape and human-made flight; Deconstructed Noir/glam; Transparent chiffon, cord, wire and wooden frames; Constructed flight/broken wings. Gold leaf. Cinematic Lighting

Alan thinks future assessments for fashion and textile design students will be based on the prompts they used to create their fashion illustrations.

“It allows us to turn our thoughts into some visual reality. That is exactly how we should use it,” he explained.

“The prompt writing for AI is the control. We are going to have to get students to reference the application they used and the prompts they used to show how much of the work is them and how much of it is the computer.

“We are entering a new phase within design education where the assessment will be ‘what did you ask the computer to do?’.

ALAN - stepford wives RESIZED

Another AI-created fashion image. This time Alan's prompts included: embroidered dress by Prada; cinematic lighting; ultra-detailed textures; photo-realistic; life-like

“Some high-profile AI users think you should never put in a designer’s name or a photographer’s name when prompting images. Is this a fast track to homogenisation?”

“If I do not use my thoughts and my knowledge, I am going to get something that looks like somebody else’s work.

“The fashion industry is very digitally driven in terms of what is in their tool box. It would be naïve to ignore it.

“Being visible with our digital skills and teaching means we are further ahead than many other universities in the country.

ALAN - white dress RESIZE

For this picture Alan's prompts included white torn shirt dresses, Margiela, Sound waves, waveforms, sea, water, wet, storm, rain, atmospheric, shipwreck, rope, wood, night, lightning, dynamic, cinematic 35mm

“We do not want to have students graduating from our course and then industry turning round and saying ‘we don’t need that’. We have to be actively current in our teaching.”

As part of his ongoing research into teaching AI, Alan is going to assign four student volunteers to use the generative AI image maker Midjourney over the upcoming school year to see how they get on with the tool.

While his knowledge of using AI in fashion textile design will be taught to students in the second and third years.

Alan also thinks AI replacing graduates is not going to happen.

“Ultimately we can spend a day using old methods to come up with a sketch and this AI can do it in an hour. It is a tool that speeds up the process for us. We are using a tool box that makes us better at our jobs.

“And we still make textiles and outfits from our designs. At the end of the day, you cannot wear a pixel!”

 

Posted on Friday 27 September 2024

  Search news archive