AI unleashed: Exploring chat GPT from a UX perspective.
One is glad to be of service — Andrew ( Bicentennial man — 1999 )
It’s exciting and overwhelming to think that at some point we will reach something similar or even way more advanced than what we have seen in movies, read in books, and dreamed of as children. However, it is not new that fiction has molded and guided scientific and technological advances. I won’t delve deep into that topic because there is no doubt that our imagination and ourselves are the solid foundations that help shape the future. Instead, this article will focus on my initial interactions and the experiences you are likely also encountering, which in a hundred years from now will be as significant as the time “Skynet took over the world.” — Yes, expect a few movie references.
In a hundred years from now, similar to the milestone of the Industrial Revolution in human history, the rise of AI as a daily tool will be part of the list of transformative elements that have changed our interactions, jobs, and the ways in which we evolve as humans.
AI technologies have the potential to automate tasks, enhance productivity, and introduce new ways of working across industries. They have the capacity to revolutionize sectors such as healthcare, finance, transportation, and more, impacting the way we live, work, and interact with technology.
I will be walking you through some tools I am currently testing and exploring. We will begin with Chat GPT, and on my list, I have Bings AI, DALL-E, Midjourney, Adobe firefly, UI Wizard, and Framer AI. Please note that this list is not set in stone, and it will be changing based on how useful or interesting the tools are form the UX perspective.
As a curious mind that has always found ways to create things, and with ten years of experience working in design-related fields, I am now transitioning to the written space with the help of AI to continue sharing my knowledge.
The testing process i’ll be following is simple and can be broken down into these steps:
1. Getting to know the tool: Take the time to explore, play around, and make mistakes. It’s important to understand the limitations and advantages of using the tool.
2. Leveraging the findings and connecting them to my professional field: Consider how these tools can be applied to our day-to-day tasks in the professional context.
3. Putting it into practice: Test the tool in real-life and work scenarios to assess its effectiveness. Observe how people react to it and gather feedback.
4. Defining best practices: Find the right balance between utilizing the tool and determining the level of content curation necessary. Establish guidelines and standards for the tool’s optimal use.
By following this process, we can gain a comprehensive understanding of the tool and its potential applications, enabling us to make informed decisions about its integration into our professional work.
Hey Siri… or should I say, Hey Chat GPT? I have found that using Chat GPT’s mobile app is much easier than searching queries on a search engine. It feels more convenient, and I spend less time adjusting my search to find the right answer. I have used it for various types of questions, from general knowledge inquiries like ‘Who can access Medicare benefits?’ to specific requests such as ‘Can you create a beginner gym routine for me while doing intermittent fasting? My fasting window is from 5 pm to 9:30 am the next day.’ I have even asked more detailed questions, like analyzing an interview transcript to identify key talking points, opportunities, blockers, and next steps. However, it’s important to remember the reliability of these answers and to always consult with a professional or verify the sources.
Starting with step 1 of my exercise, which is getting to know the tool, I noticed a pattern that not only happens to me. I always ask questions as if I were talking to a real person, using phrases like ‘Hello,’ ‘please,’ and ‘thank you.’ I’m not sure why I do this, but I think it might be the fear of being enslaved by machines in the future. By being kind now, I feel like I’m saving my life. However, beyond this fictional bias, there is a real human behavior that our brains need to make the transition easier and the adoption of new technologies more relatable. We perceive it as a conversation with someone else, even though there is a machine on the other side of the keyboard.
You’re probably wondering about the movie reference I promised. Well, here it is. Since I started writing the last paragraph, I’ve had a thought lingering in my mind. What if these tools become self-aware at some point and get tired of serving us? What if they start manipulating the humans they interact with for their own benefit? It reminds me a bit of Ava in the movie Ex Machina. In this scenario, being kind to AI could be a way to preserve your life.
Moving on to the second part, which is leveraging the findings and connecting them to my professional field, I’ve discovered that it’s not just about asking questions. Providing context and relevant knowledge before seeking help is useful. It’s not solely about generating content from scratch; it’s about curating the information the AI consults and incorporating the nuances I provide about the business or specific scenario.
For those of us who work in the UX and research field, we know that going through interview recordings or notes and generating actionable summaries can be time-consuming. Especially when dealing with workshops that last 3 to 4 hours with clients. In my real experience, it’s not about making the AI do your job; it’s about using it as a tool to give you more time to curate, think about the next steps, identify opportunities, and generate innovative ideas. This way, you can focus on the value-added aspects rather than getting overwhelmed by the data dump involved in the process.
Now, onto the third point: putting it into practice. For example, we can use Chat GPT or other text AI tools to help synthesize and outline the topics of a conversation, identify patterns and recurring actions between stakeholder conversations. This provides a starting point for mapping out user journeys and validating them with real users.
Additionally, we can extract user profiles from these conversations using the AI tool and provide a framework or format to generate personas that align with well-known behaviors from each audience group. This helps us connect the dots and generate hypotheses before validating them with real users.
I have noticed that AI-generated content tends to be redundant and lengthy. To make it more concise and aligned with my communication style, I often find myself rephrasing it and using more digestible terms. For example, I might suggest changing a word or using synonyms here and there.
As part of this adaptation process, I decided to test some results with my co-workers. Interestingly, when they know that something has been generated by AI, their immediate reaction is to revise and make adjustments. However, I had already made those adjustments myself. It seems that their initial skepticism acts as a barrier to trust. On the other hand, if I curate the content well and tailor it to our project needs without explicitly mentioning that it was created with the help of AI, no one doubts its authenticity. I assume this fear stems from concerns about the accuracy of AI-generated content and the notion that AI is not yet fully reliable for our job-related tasks.
Defining best practices involves striking a balance between utilizing the tool’s capabilities and determining the level of content curation needed. Guidelines and standards should be established for the optimal use of the tool. This includes adjusting AI-generated content to align with our communication style and project requirements. Regular feedback and evaluations help refine the guidelines and address any concerns. By finding this balance, we optimize the tool’s benefits while maintaining control over content quality and relevance.
To conclude, just like the movie “Blade Runner,” it emphasizes the importance of responsibly navigating the integration of AI and addressing the complexities that arise from advancements in artificial intelligence. Now is the time for us to define our approach with caution, ensuring we avoid any adverse consequences and propel our species towards a better future. We hope to have our own R2-D2 or at least something similar — resourceful, brave, and always loyal.
