How might we use Artificial Intelligence to help MA Design Management students learn to create effective visual narratives?

Also:

• to begin to integrate the potential of AI into their practises,
• to embrace the potential of AI as a design tool
• to recognise and avoid the potential of homogenisation of design strategy through use of AI.

A first reflection

“…Such work will be considered as malpractice, as the words generated are not the students’ own. The only exception to this, is where the course content permits the use of AI generated work.” (UAL, 2023)

The current approach to student use of AI at UAL is ambiguous, including both the fearsome language of ‘malpractice’ as well as the potential for an exception with permission. In addition to the “growing unease around whether the behaviour of these systems can be rendered transparent, explainable, unbiased, and accountable”  (Shadbolt, 2022), design educators in their hearts could be struggling with existential fears: of a great homogenisation of our work and appetites, and beyond that, of the replicability of the unique humanity of our creativity.

“Concerns about its unintended effects and misuses have become paramount in conversations about the successful integration of AI in society.” (Manyika, 2022)

To be fair, UAL is currently engaged in pilots to create turnitin plug-ins to identify use of AI tools(so they can be identified with appropriate references), as well as the development of new policies to enable the ethical, equal and transparent use of AI in students’ work. For example, consideration is being given to authorising and licensing specific AI tools in order to ensure fair access for all students to equivalent tool capabilities. (Elena Hernandez-Martin, Jan. 8, 2024)

But reality speaks: design businesses are using AI as a tool to make traditional activities more efficient and clients are asking their design partners for their expert advice on its use. (Nick Vaus, November 2023)

References

Dewit, I., Rohaert, S., Corradi, D. (2021), ‘How can comparative judgement become an effective means toward providing clear formative feedback to students to improve their learning process during their product-service-system design project?’, International Journal of Technology and Design Education 26(3):276-293, November 2021. Available at: https://www.researchgate.net/publication/356036811_How_can_comparative_judgement_become_an_effective_means_toward_providing_clear_formative_feedback_to_students_to_improve_their_learning_process_during_their_product-service-system_design_project

Heneberry, P. & Turner, A. (2016) ‘Critical action learning –
rituals and reflective spaces’, Action Learning: Research and Practice, 13:1, 60-68, DOI:10.1080/14767333.2015.1130349
Available at: https://doi.org/10.1080/14767333.2015.1130349

Manyika, J.M., ed. (2022) ‘AI & Society’, Daedalus, Spring 2022. Available at: https://www.amacad.org/daedalus/ai-society (Accessed: Jan. 2024).

Nixon, N., (2023) ‘This is the AI we didn’t see coming: artificial imagination‘, Fast Company, 04/26/23. Available at: https://www.fastcompany.com/90887178/ai-we-didnt-see-coming-artificial-imagination. (Accessed Jan. 30, 2024).

Shadbolt, N. (2022) ‘So Simple a Beginning: Species of Artificial Intelligence’, Daedalus, Spring 2022. Available at: https://www.amacad.org/publication/so-simple-beginning-species-artificial-intelligence (Accessed. Jan. 2024).

Sharma, R., Kumar, S., Pathak, K. (2023), ‘Technology and Society’. Available at: https://www.researchgate.net/publication/376516453_Technology_and_Society [accessed Jan 28 2024].

Silver, H.F. (2010), Compare and Contrast: teaching comparative thinking to strengthen student learning. Publisher: Association for Supervision and Curriculum Development, Alexandria, Va.

UAL (June 2023) Impact of Generative Artificial Intelligence tools on Higher Arts and Design Education Available for download at: https://lccteaching.myblog.arts.ac.uk/digital-champion-ais-impact-student-learning-teaching-and-assessment-in-arts-and-design-higher-education/ (Accessed Jan.30, 2024).

UAL (Feb. 2023) UAL Awarding Body. Available at: https://www.arts.ac.uk/partnerships/ual-awarding-body/stories/chatgpt,-ai-and-ual-awarding-body-qualifications  (Accessed: 25 Jan. 2024).

Ethical Enquiry

The aim of this enquiry is to begin to understand how generative AI can be used to help students learn to develop effective visual strategies. It is also intended to begin to integrate the potential of AI into their practises. To embrace the potential of AI as a design tool, and to recognise and avoid the potential of homogenisation of design strategy through use of AI.

The 2023-24 cohort of the MADM (29 students) will participate in a working session as part of the curriculum.  Prior, they will be contacted via their email addresses with information about the study and a link and/or return slip to inform the researcher that they agree to be included in the work session. For those people who fail to return the form, an online form will be shared on the day of the session (December 4, 2023). All respondents will be contacted via email the following week to gather their post session reflections. Students will be given the option to opt out at any time, and simply participate in the learning experience, which takes priority over the research. However, as this work session is part of the research unit, it is expected that the transparency of the research process and consent will provide a real-life research experience that students can use in their own course research requirements. Sam Barber will be a co-leader of the working session and will be asked for her reflections as well.

Research Methods

This project utilised the Action Research Project Framework as a means of assessing and improving my educational practise during an active teaching process. It’s phases include Observing, Reflecting, Planning and Acting, after which a Reflectionm stage prepares for the next, improved iteration of the teaching process. Learning by doing.

Further rationales for the specific research methods chosen will be offered as introduction to each of the four activities of observation, reflection, planning and action.

The research activities included, by phase:

1. Observation:

• Understanding the current state of the relationship between visual strategy and AI in a London agency. Expert interviews: Nick Vaus, Founder, Free the Birds, David Clancy-Smith, Design Director & AI Design Leader at Marks Examples of work, quotes.
• A brief exploration of alternative visual AI tools
• Back-casting an existing visual strategy to establish a point of view on visual AI limitations and strengths
• Understanding UAL policy: Interview with Elena Hernandez, Digital Leaning Producer, Digital Learning, UAL….

2. Reflection

• Additional secondary research: (Artificial Intelligence versus Artificial Imagination)
• Consideration of learning methods: co-creation, comparative judgement

3. Plan of Action

• 4 hour student work-session to develop visual strategy capability, gather their POV: how to teach it AND suggestions for future approaches
• Developing the lecture material

4. Action & Data Collection

• Delivering the work session
• Personal observations
• Gathering student feedback
• Workshop visual outputs
• Peer observations (Sam Barber, co-lecturer)

Part 1: Observing the current situation

The double edged sword of technology in design leadership

Activities:

• Understanding the current state of the relationship between visual strategy and AI in a London agency. Expert interviews: Nick Vaus, Founder, Free the Birds, Dave Clancy-Smith, Design Director & AI Design Leader at Marks
• A brief exploration of alternative visual AI tools
• Understanding UAL policy: Interview with Elena Hernandez, Digital Learning Producer, Digital Learning, UAL

Learnings

CONTEXT & Situation: critical thinking for the MA Deisgn management student

This research is being done in the context of the learning program of students in the MA Design Management. Many of the students have undergraduate degrees in design, engineering, or some other creative field. They come to the course with the hope that the curriculum will elevate their careers into roles with influence to shape the interaction between ‘clients’ and designers in a way that results in powerful, satisfying outcomes. Our students are expected to be leading design that creates meaningful, positive change in the world at a point in which society in struggling to find a way to balance the needs of people and the planet.

Critical thinking is essential in the development of the creativity behind design/ideas that change(s) the world in meaningful ways.

Strategic thinking is not just about the written and verbal story; the use of a multi-sensory storytelling toolkit is essential to a powerful design strategy. This small intervention is intended to explore the teaching of the use of AI in the development of visual ideas/strategies that are differentiating, interesting and powerful in their emotional and aesthetic impact.

Technology, society & design: dramatic impacts

AI is one of the game changing technologies that will inevitably alter the way we live and work as a society, impacting daily life not only in practical ways, but also with both positive and negative societal shifts.

Sharma uses the case of social media’s influence on civilisation to point out that this technology based innovation has exposed society to false information, echo chambers, and cyberbullying even while giving people the benefits of access to information, connections across time and space, and the ability to mobilise for social concerns and raise their voices. (Sharma, 2023)

Closer to home, Adobe’s photoshop has had a huge impact on the work of graphic designers and designers: but, as a Creative Director from Saatchi once said to me, “It hasn’t really created better ideas, but has given clients the expectation that they can get them faster”.

Potential for inequities: not all students will have access (financial) to the latest technologies. (Sharma 2023), (Hernandez, 2023)

While UAL’s published policy regarding students’ use of AI is at once fearsome and ambiguous, policies are being developed in response to the pragmatic appetite for its benefits from students, tutors and industry.

Current recommendations (UAL, Impact of Generative AI tools, 2023) include

• Identify opportunities where AI tools can enhance learning through practice and provide rapid feedback to students
• How to use AI to inform parts of the creative process rather than guiding it entirely
• Incorporate workshop sessions to educate students on working with different AI tools, including understanding their limitations
• Inviting industry experts relevant to the course
• Facilitate staff training in the tools

AI is already in use in commercial design

“It takes us 1 day instead of 5 to source the images to bring an idea to life”

(Vaus, 2023)

Agencies are using both verbal and visual generative AI . Generally, it is an interaction between MJ and ChatGBT, combined with Photoshop. Int he case of both interviews, ChatGBT (verbal) and Midjourney (visual) were the preferred tools.

“We’re using AI as a sharpening tool for language. For me (dyslexic), it makes my writing more effective: brainstorming headlines, shaping language, the verbal side of thinking and as a quick shorthand for visualisation rather than sketching;

…so much quicker and faster, concept boards but also conceptual packaging/deliverables/interiors/POS”

Nick Vaus, Founder, Free the Birds

CPG clients are also asking for opinion about the correct use of AI in their branding identity and packaging deliverables. Clients are inquisitive about this; marketing teams within these organisations want to know how they can tap into this…we’re the educators at the moment

Free the Birds product concept exploratory:

“It’s nothing without an idea”

David Clancy-Smith, Design Director & AI Design Leader, Marks

AGENCies are using midjourney to create efficiencies

• has reduced the investment by designers in the development of concept boards/visual strategy boards from one week to one day. Designers can develop alternative strategic expressions very quickly.
• We can visualise things super quick…ideas into product development a lot faster
• Should enable exploration/development more ideas because sometimes limited by time, budgets…should allow a breadth of ideas spin through
• Allow more ideas to be tested, more quickly (overnight digital testing)
• Ultimately shortening time to market…more freedom fo

Watch-outs:

• Owner intellectual property
• • Need to be constantly aware that this won’t create ideas for them…only a tool…if you do rely on it (designers/clients) the answers will be built on current reality…so answers aren’t there…you still need creativity: ingredients to create something that’s never been seen before…with Midjourney, because its never been seen before, it cannot do e.g. Hands on a clock cannot be turned to spoons
• • So you need to take it into photoshop for the spoons for hands…power is always with the creative.
• • Answers for ChatGBT…can ask ChatGBT questions and get answers..can’t ask Midjourney questions…constantly feeding mj with what you want it to build.
• • Midjourney requires inspiration..quite a different relationship

Dave:

• Often quite topline,
• • Quick ideation
• • Not great at original thinking
• • Is nothing without an idea
• • The Midjourney aesthetic, chaos

•

a bRieF ai exploratory

There are a number of visual AI tools for design: https://www.unite.ai/10-best-ai-graphic-design-tools/. But this intervention focusses on the emerging area of image generations which has also historically been a central activity for graphic design firms and in the area of their strategic narratives. https://zapier.com/blog/best-ai-image-generator/.

The following mini exploration is a bit of a back cast from a prior visual strategy from my own practice: the creation of a brand strategy for a Russian vodka company intending to introduce itself to global markets.

Generated by Adobe’s (Free) Firefly: “people drinking vodka”

I wasn’t incredibly impressed. So I modified my prompt:

Not much better, really.

COntrast: Using Midjourney to back cast an existing strategy

Some context: During the pandemic, while everyone was working remotely, I worked with a Russian spirits company, Ladoga, to develop an international brand from their market leading super premium vodka in Russia, Tsarskaya Gold. The Discovery involved understanding the current brand strengths with Russian people, exploring the global competitive context, and looking for emerging opportunities to provide a differentiated offer in the international market.

This case is relevant to the students because it was about finding an idea that was leveraging existing brand equities/strengths, but also moving competitively into the future market.

The brand strategy was a composition of writing and visuals intended to communicate not just performance/functional attributes, but also provide an emotional and experiential design brief for product designers, marketers, communicators to inform all designed experiences with the brand.

The Tsarskaya Gold brand strategy

Images provided by Midjourney using the strategic ideas

The designed brand narrative using the MJ images

“As a collection of ideas adding up to a whole, it lacked authenticity (it wasn’t even Peter the Great!). It also lacked the balance and tension of a good idea, in this case, between the past and a present Russia that was different from the cliches in the MJ image. And it lacked the humanity required of great designed experiences.”

Crombie, 2023

“the MJ version failed to communicate the ideas that I wanted, even though it looked polished and impressive to other people”

(Crombie, 2023)

Auto-ethnography: When I showed this result in the lecture, both the students and the peer observer responded well to the new, MJ inspired visual strategy. They thought it spoke well to the strategic words (top of mind, as well as the compared to the original). But I was disappointed in both their response and in the outcome.

As a collection of ideas adding up to a whole, it lacked authenticity (it wasn’t even Peter the Great!). It also lacked the balance and tension of a good idea, in this case, between the past and a present Russia that was different from the cliches in the MJ image. And it lacked the humanity required of great designed experiences.

Part 2: Reflections and further enquiry

Insight from Observations recap:

• Digital technology will and has had dramatic impact on the way we live and work, for good and for bad
• Potential for inequities: not all students will have access (financial) to the latest technologies.
• AI is already in use in commercial design agencies and is a source of real efficiencies; their clients are asking for it.
• But visual AI tools still require an idea to result in strong design work.
• And the current design agency tool of choice, Midjourney, has weaknesses when communicating ideas/strategy: authenticity, humanity, subtle layered tensions
• UAL’s policy (currently ambiguous) is in development, but recommends finding opportunities for generative AI to inform parts of the creative process, encourage workshops in which students can learn how to use them, but also their limitations….

Reflection: ultimately the teaching objective for the work session is to help students learn how to translate verbal ideas into strong visual ideas using available technology tools in the most effective, efficient way possible. This includes an understanding of the potential of AI, but also its limitations.

1. The potential of a comparative judgement approach to learning & assessment (Dewit, I. 2021)

• Comparative judgement can be an effective teaching tool in assessing difficult constructs like a product design process: students can experience and assess for themselves the differences in effectiveness between one approach/tool/outcome and another.
• Also is more focussed on the holistic project result, using frequent/multiple feedback events (and contributors) to develop a sophisticated, critical understanding
• The idea is to ask them to assess the effectiveness of the AI tools in producing images that communicate their ideas in order to build a critical understanding of how to develop effective visual ideas)
• They can also learn the strengths and weaknesses of the tools for their own practise

2. We will build upon the ideas they have been developing throughout the unit rather than creating new ones

• The students have been organically and iteratively developing ideas and insights throughout their Research Unit; the intention is not to create new ideas, but to be begin to translate these well researched and nuanced ideas into strong visual/sensorial designs.
• They will use the insights from the Assignment they’e been woking on though the Unit: to identify a problem area/design opportunity based on one of accessibility, culture, pollution and retail in E&C
• The work will continue in the Design Leadership Unit by inviting Nick Vaus, founder of Free the birds, to help them use AI to develop visual concepts from ideas.
• Artificial Intelligence versus Artificial Imagination

3. Four ways to use generative AI as a co-creator: (Is our imagination under threat with this injection of generative artificial intelligence?) (Nixon, 2023)

1. AI as a collaborator to spark our imagination. Generative AI as a starting point, not the outcome
2. AI as a catalyst to help us learn to ask better questions. For Midjourney in particular, it requires an inspired, informed prompt.
3. Background research and integrity matter: AI can only source from existing material and so are not creating new, ownable insights, but compositions of the current (and other people’s thinking)
4. Critical thinking requires multiple perspectives; and the tool is only one.

Part 3: Creating the Plan of Action

The Action Research Event:

• 4 hour student work-session as part of the Design Research and Methods (DMRM) Unit of the MA Design Management Course.
• Intended to develop students’ capability to visualise ideas in a compelling, effective way.
• Reveal an opportunity for generative AI to inform parts of the creative process
• A workshop in which students can learn how to use visual AI tools, but also realise and co-define their limitations.
• Using a comparative judgement approach to develop their critical point of view

The Agenda:

Part 1: Co-defining visual strategy: cases, homework, semiotics intro, exercise

Part 2: the work session:

1. In teams based upon students’ subject categories to create a strategic board to set up the context for your subject category
2. Align on your top 4 insights
3. Enter these insights into Midjourney to create a board that tells the story of your subject
4. Then use them to search google for 4 images to tell the story
5. Enter all of your compositions on miro
6. Share and Discuss: (What’s good, bad, missing)

Work session presentation

https://artslondon-my.sharepoint.com/:p:/g/personal/l_crombie_lcc_arts_ac_uk/EbwdJUYWqs5KhXCvt6M7wwgBmW2-3Livw-_ohLcRjqwBPg?e=ql7pFi

Part 4: collecting data and reporting results

https://miro.com/app/board/uXjVN0E1iZM=/?share_link_id=240757480120

The Work session

“Mid journeys interpretation made it fantastical, and it’s not necessarily placed in there. It’s almost overemphasising it while we’re trying to show an under representation. It’s a made-up understanding.”

student 2023

“And the Google images sort of helped us connect with the environment and connect with the area a lot more and that’s what I feel surprisingly in the second Google image that we have, we just zoomed in there’s a lady sitting over there And we actually interviewed her.”

student 2023

“Our community is about people and it’s about creating a sense of damage and the time it took place. And so I think when you feel like people who represent the community are represented in the images you’re just you’re tied into feeling There’s a there’s an emotional connection you could make there.

student 2023

“I think the traffic and the homelessness on the top, the top two images, they are generally up to what we searched on Google as well, although it doesn’t really show the problem… I mean it does look like a heavy traffic, but it doesn’t really entail the fact that it’s causing any issues or un-safety and the homelessness is like the the tents are very neatly.”

student, 2023

“on the Google search, I think we were trying to find more of the realistic side”

student 2023

“But what you bring to that in your head and the nuance behind that insight is you are looking for something else. Whereas the AI is only looking for those words…very surface”

student 2023

‘You can change your prompt 20 times and you still won’t get what you want.”

student 2023

Insights and Findings

Auto-ethnography

• by comparing the outcomes, students became increasingly aware of the intentions and possibilities of the visuals.
• They thought Midjourney in general, was capable of creating quite futuristic and fantastic imagery from their prompts
• but that it lacked authenticity, emotion and nuance that they seemed to be able to better find in google
• they became increasingly self aware of the intentions of their ideas and how the tools were achieving them or not
• some students already have a Midjourney subscription, and their facility with it was better than others’: we need to make sure everyone has equal access to the tools

Peer observer (Sam Barber)

• High levels of engagement and curiosity on the format and content of the teaching session from students.  Evidenced by the energy in the room and the quality of discussion within the groups 
• the comparative nature of the teaching session stimulated conversation and dialogue
• Some students have access to the AI (mid-journey) already which indicates an existing level of interest in the tool and its use, there may be some capability and they are already using it.  Students are ‘hungry’ for teaching methods and discussions around the topic of AI and its use in design contexts.  UAL feels behind the curve on this issue.
• The intervention created learning on several levels – criticality on AI and its limitations and opportunities for design and in educational use for the course, the use and importance of imagery in narrative development, and the opportunities to use more creative methods in design research.  
• Using an artefact/tool such as mid-journey can create a focal point for multiple levels of learning 

what worked:

• the intervention was tailored to an assignment they were working on 
• energy in the room was great 
• setting the pre-work of coming up with insights from their assignment homework – created a deadline for secondary research to be synthesised so secondary benefits of creating momentum for the assignment

opportunities:

• Ensuring technology is more widely available within the groups and the set-up of the tech, I think this was partly a time-available issue. Whilst it worked ok in the session, I can see how if we repeated it, we could make that element run more smoothly
• More time would have been great, I can see the opportunity for a full-day workshop combining it with other elements of visual communication and visual narratives

Presentation

https://artslondon-my.sharepoint.com/:p:/g/personal/l_crombie_lcc_arts_ac_uk/EdNvL_JCdNNHt1ZvLV4vaz0BbivTOVVcUs0qCyyXtm-FBQ?e=tVVTFy

Consent forms