AI meets UX: Designing a human-centric, AI-powered authoring experience

ChatGPT set the record for the fastest app to reach 100 million users, reaching the milestone just two months after launch (a record that’s since been broken again by Threads).

With numbers like that, it’s clear that a lot of people worldwide are excited about the possibilities posed by generative AI—including Learnosity. In fact, we’re deep into development of an upcoming AI-assisted authoring tool, which we’ve named Author Aide (sign up for updates on that product here 😉).

However, as with all paradigm-shifting innovations, some users feel some trepidation about working with this unfamiliar and sensationalized technology. 

Because Learnosity believes AI is a force for good, we want to deliver a product that’s impressive enough to win over even the staunchest of skeptics. To achieve that goal, we need to design a first-rate experience that guides users into this brave new world with confidence.

That’s where our Senior Product Designer Daleen Rabe comes in. We asked Daleen to tell us about the importance of UX when dealing with AI.

Can you briefly explain your role in this AI project?

As the designer, my role is quite layered. Some of these layers include:

• Considering the user in every solution, idea, and discussion. This can include both providing value and making the product as understandable and intuitive as possible.
• Addressing the ethical aspects of AI technology by focusing on transparency, user rights, and values.
• Adding a user-friendly aspect to an artificially generated experience. 
• Incorporating prompts visually into designs to gently hand hold users into a new authoring experience and help them obtain the best possible output.
• Facilitating discussions to align a broad range of stakeholder views to create a unified vision.
• Practicing fast iterative design processes, prototyping, and testing that grows with the team’s knowledge and the AI models we leverage.
• Gather feedback, conduct usability testing, and gain insights to further improve the designs.
• Creating interactions that balance the technological capabilities with both the user and business needs in mind.
• Creating solutions that can scale based on the team vision and user needs.

What was your first reaction when you found out that we were going to integrate GPT-4 into our product?

From a design point of view, my first reaction was a combination of excitement and nervousness. When it comes to AI, there’s a vast ocean of knowledge and undiscovered potential. It felt like staring at a blank canvas, but knowing that this can be something great.

What did your research look like for this project? 

As this space is new and in constant development, it’s unlike a “normal” design process. Not only does the actual AI model continuously change, so does our own understanding of what’s possible. For this reason we see our product as continuously evolving and we’re approaching it as such. We’re currently developing a beta that’ll serve as an experiment and will mainly be used to gain more insights into our future development plans. 

Using past experience and interviews with various customers, we knew the general direction that we could take. We’re also in continuous conversations with existing customers and their feedback forms an integral part in how we approach the designs.

From a technical point of view, the team itself took various means of gaining the necessary understanding of this space and technology. I took up a short course in the model we investigated and reached out to my network to gain more insight in what the technology can offer.

Our team of designers also joined a meet-up, read articles, and generally scratched that itch of curiosity around what could be possible in this space, and how we can leverage it to deliver value to our end users.

As designers, we view this as an excellent opportunity to embark on our continuous learning journey and broaden our understanding of the data realm and human-computer interaction in general.

What user experience were you trying to achieve?

For the first iteration, we wanted to make it understandable with minimal intervention so that users can experiment with some basic tasks independent of our guidance. For this reason we kept the interactions familiar and used language that’s friendly but informative. We also used techniques such as chunking, visual anchoring and delighters to create a simpler and more enjoyable experience.

Through this first iteration we also want to gain more insight into what our users need and how we might improve the experience overall. We achieve this by experimenting with discoverability, decision fatigue, and existing mental models.

Why should content authors be excited about our AI-assisted authoring tool?

We’re just seeing the tip of the iceberg of what AI can do for authors right now. With our beta release the experience is basic by design, but it can already offer customers the ability to create content faster and more efficiently.

It’s not only a time saver but a creativity tool. We’ve all had those moments of thinking: “where do I begin!?” This tool assists authors in overcoming the blank canvas effect or experiment with some more unknown areas of knowledge.

What were the major challenges and achievements of this project from a designer POV?

The main challenge is managing the unknown and the fast-paced nature of this space. Designers often deal with ambiguous situations, but this project is particularly ambiguous. With AI, the future is uncertain, the pathway is unclear, and the expectations are in the process of being discovered.

From a design perspective, the challenge is to be comfortable in the uncertainty, find the pathway forward in creative ways, and stay curious about the users’ expectations. 

In terms of achievement, it’s incredible to watch our technology and tools work together with AI to drive value for our customers. We’re making real progress here. 

What advice would you give to other product designers looking to incorporate AI into their user experiences? 

Be open to change, listen to a variety of perspectives, and ask loads of questions. It’s a daunting space, but Rome wasn’t built in a day. Identify your knowledge gaps and prioritize those based on the project at hand. If you don’t know certain terms or you’re lacking a specific area of knowledge, write them down and investigate. 

Here are some tips and resources to get you started:

1. Take advice from trailblazers in this space, like IBM and Microsoft.
2. Do online courses based on your project needs and knowledge gaps. For example, a short course in ChatGPT basics, IBM’s artificial intelligence practitioner course, or Google’s learning path on generative AI.
3. Reach out to your own network for guidance and help on the topic.
4. Keep learning and be curious to find your own resources.
5. Stay up to date by following thought leaders in the field on LinkedIn.
6. Go easy on yourself, it’s a complex domain.

In your opinion, how will AI shape the future of assessment?

AI will transform the assessment landscape completely. It has the potential to change the way we create, deliver, and analyze the full learning journey. The possibilities are only limited to how far companies are willing to integrate the technology.

As designers, it’s our responsibility to ensure that the holistic human experience is considered and this is no different when it comes to designing with AI in mind.

I believe that true success in technology is when it improves human life—and designers strive to achieve this in every decision we make. 

In the lead-in to our AI product launch, we’ll be sharing related content and updates. Sign up below to stay connected 👇

Learn all about our mission, vision, and manifesto for AI—download our AI guidebook for deeper insights on what AI means for assessment. 👇