Using Artificial Intelligence to Improve a Digital Product’s User Experience
Artificial Intelligence has been a hot topic for quite some time now.
Countless blog posts, newspaper articles, and tweets warn us about the imminent end of the work place as we know it or remind us of essential AI skills we need to catch up with the competition or to skyrocket our productivity.
Image compositions generated by Artificial Intelligences like Midjourney, Dall-e2, or Firefly are impressive and, especially for us designers, alarmingly good.
Undoubtedly, the current developments in the field of Artificial Intelligence are groundbreaking and will change our world profoundly.
Large Language Models: Behind the Hype of “AI”
Artificial Intelligence is not easy to explain. A precise definition of what “intelligence” means, would exceed the scope of this article. Other authors have also pondered this definition and found agreement with something like this:
“In other words, automated tasks that require interpretation, combinatorics, and non-trivial execution can be considered Artificial Intelligence.”
From “A Very Gentle Introduction to Large Language Models Without the Hype” Mark Riedl
A commonly cited example of Artificial Intelligence can be found in video games: Game characters that are not controlled by human players but react to the player’s actions in the game. Simplified, this behaviour is achieved through numerous if-then statements (“if the player is within range then shoot else move to the nearest boulder for cover”), which are sufficient to create the impression of a “living and thinking” entity interacting with us.
Artificial Intelligence relies on vast amounts of data. In a “Large Language Model,” these data amounts serve as the “knowledge base” for algorithms capable of extracting patterns from them. It can be compared to a search engine like Google’s index, but on a massive scale: an unimaginable amount of information gathered through systematic web scraping – useful, factually correct, but often also useless and erroneous.
Significant improvemens have been made with the ability to interpret natural language inputs and provide seemingly appropriate responses through pattern recognition. But whether an answer is considered correct is essentially the result of probability evaluations enriched during the model’s training.
Use Cases for AI in Creative Industries
How can Artificial Intelligence support us? What are the fields of application for Artificial Intelligence in the daily work of a creative?
The strengths of Artificial Intelligence lie in these areas:
- Pattern recognition
- Image recognition
- Natural language processing
- Automated data processing
- Learning ability
- Refinement and processing of unstructured data
There are numerous tasks in which we already greatly benefit from the use of Artificial Intelligence:
- Automated Content Creation – Text, Graphics, Images & Videos
Artificial Intelligence can be used to create text, graphics, images, and videos. Especially during brainstorming and mood phases, AI opens up new possibilities to quickly achieve initial results and generate output. This does not replace “classic” content production but rather extends and accelerates the process. - Automation of Standard Tasks in Image Editing and Motion Design
A significant reduction in working time for tasks like image enlargement, retouching, color adjustments, and similar tasks. - Artificial voiceovers and soundtracks
In the field of motion design, Artificial Intelligence can be used for quickly creating voiceover commentary or music accompaniment for video content. - Research Work, Support in Problem Solving
With all the caution required when using Artificial Intelligence as a “knowledge provider,” it helps to quickly, efficiently, and precisely expand knowledge. - Structuring and Optimization or Translation of Text Data
Artificial Intelligence can handle stylistic text tasks just by placing a few text prompts: shortening texts, style adjustments, structuring lists, or translating into other languages. These tasks don’t always provide the perfect result immediately, but the efficiency boost from using Artificial Intelligence is remarkable.
So far for using standalone Artificial Intelligence solutions for support work.
But now, let’s take a look at a custom Artificial Intelligence integration I’ve been working on at dmcgroup in order to improve the User Experience of a software tool…
“Situation R®”: Software for Conducting Decision-Making Workshops
Our client FAS Research is a leading company in network-based stakeholder mapping and strategy design. In a shared team led by FAS Research, we have been developing the software for one of their consulting products for several years.
FAS Research enables teams to make significant strategic decisions in a short time, applying interactive group processes. This happens in collaborative workshop formats (both on-site and remote), where participants vote on questions developed collaboratively during the workshop, leading to a consolidated situational assessment on complex topics – all within just a few hours.
The software we have been developing in collaboration with our client FAS Research supports both their moderators in preparing and conducting workshops as well as workshop participants in voting using their smartphones.
Reducing Cognitive Load for Moderators through Artificial Intelligence
The work of FAS Research moderators is challenging: during the workshop, topics for which a vote is to be taken are usually worked out “live” with the participants. Meanwhile, moderators must moderate the process and simultaneously record, consolidate, and build consensus on the suggestions collected in the group. A cognitively demanding multitasking activity.
Moderators play a significant role in the success of FAS Research’s workshops, thanks to their ability to actively support groups in the decision-making process.
The tasks involved in recording the participant’s inputs are well suited to be supported by Artificial Intelligence:
- Summarizing/shortening a list of text elements
- Expanding/completing a list of text elements
- Adapting the text style (e.g., formal, informal…)
- Translating text into other languages
For some time, FAS Research has been planning to integrate Artificial Intelligence into the workshop software and found a reliable and creative partner in dmcgroup to implement it.
Our joint solution in the moderation software relies on the integration of the text AI ChatGPT. Using a prompt (a type of command that gives Artificial Intelligence instructions for processing the provided texts), existing texts are passed on and converted to the desired format in no time.
Nevertheless, moderators always have full control over which text suggestions are accepted or rejected.
However, by using Artificial Intelligence to handle time-consuming routine tasks, moderators can save valuable time and fully focus on moderating the workshop and collaborating with workshop participants.
Using Artificial Intelligence to create content from scratch
Text is not always collaboratively developed in the workshop. Sometimes, moderators prepare are workshop before the actual session.
In this case, AI integration enhances efficiency: with just a prompt, a list of elements can be created “from scratch.” This allows stylistic work and research effort to be outsourced.
Our latest software feature update uses “Whisper,” an audio AI that directly translates audio recordings into text.
Using this feature in the workshop, AI automatically captures what is said in the workshop in text form by transcribing an audio recording of the session. Of course, this feature will only be used with the prior consent of all participants and in compliance with all data protection regulations.
Moderators can use appropriate prompts to further process the automatically transcribed content – for example by consolidating it into a concise list directly.
Iterative Improvement – AI always in the Loop as a “Sparring Partner”
Iteration of text inputs is at the core of AI integration in the Situation R® software.
No matter how the “AI journey” begins in Situation R®:
- Generating text “from scratch”
- Working further existing texts
- Transcribing an audio recording
The use of AI does not end here. Throughout the work, the context of previous text manipulations is always “remembered” by the integrated AI. Each new prompt builds on previous commands and content, allowing for the iterative improvement.
The seamless integration of Artificial Intelligence into the software eliminates the need to switch between browsers or tools. Moderators work step by step on creating and optimizing text content, just like in a conversation.
Handle with Care: Drawbacks of Artificial Intelligence
Data protection and privacy of customer information are of the utmost importance to FAS Research. This is an essential duty in the relatively young field of practical AI application.
Like with any technology, some “dark sides” should be considered:
- Artificial Intelligence is ultimately a black box
The underlying algorithms, knowledge base, and decisions remain in the dark. Results from AI can vary from one query to the next. Answers carry a certain degree of unreliability. - Data Privacy
Every query to Artificial Intelligences like ChatGPT is stored on external servers – and may even be used for training AI. It goes without saying that confidential data should not be fed to AI. - Fake Information
The results provided by Artificial Intelligences are composed of found internet data selected as responses based on statistical models. Since this is purely probabilistic, there is no dimension of true and false. Often, AI provides factually incorrect answers.
Like any technology, the use of AI carries potential drawbacks.
However, with the necessary sensitivity and control, its use offers exciting new possibilities for improving the user experience.
This article was originally published for dmcgroup.eu in german.