Surviving AI Gold Rush: Shifting Focus from "UX with AI" to "UX for AI"

This is a human-written article about User Experience for Artificial Intelligence Technologies.

After ChatGPT, Most people in the UX world are more concerned about incorporating AI in UX that is optimized. Initially, even I was excited about AI applications in UX. But fortunately, I had the opportunity to get my hand dirty working on projects related to exploring AI. Then only I realized that As UX professionals, we should embrace the perspective of “Seeing AI as technology and How can we distribute this technology to users with good user experience?” rather than emphasizing more on “AI as a competitor or complement to our Individual selfs”

Why ChatGPT is starting point of the new UX Revolution?

SourceIn fact, everyone thinks ChatGPT is only AI Revolution but it's more to do with UX. ChatGPT is an abbreviation of “Chat Generative Pre-Trained Transformer” The first paper academic paper introducing Transformers technology was published six years ago by a team at Google, if you were really excited by these capabilities by Transformers a year ago you actually excited by AI capabilities. if you’re getting excited now, then you are getting excited by UX.

If you are using Netflix or Google for the last few years then you are already interacting with AI technologies because for almost around for the past 60 years, the UX approach is seamlessly delivering a good experience to the user and don’t make them worry about what’s actually happening behind the screens.

Basically, the product builders want the User to click the right buttons at the right time in an already pre-determined User Interface designed by them. They don't want users to know what is the technology behind them, whether is it deterministic instructions given to computers via programming commands or probabilistic smart recommendations given by Machine Learning models. All we cared about is delivering a seamless intuitive user experience with User-Centered Interface backed by research than guess.

This is the first time with ChatGPT and later Bard. The product builders want users to know about the technology behind the screens. And allowing users to play with AI using the Prompt-driven User Interface. If you deeply think populistic products like Facebook, Instagram, and youtube had a specific purpose like social networking, and entertainment for adopting it. But it looks like ChatGPT doesn't have a specific purpose and it's open-ended and is almost the first time in user experience that users dont have a specific pre-defined purpose they can use it subjectively as they want by using a user interface, In which users can send input by articulating what they want as output in written format and it will bring back out something relevant to it. Problems UX professionals need to Solve

1. Prompt Driven User Interface Hurts Usablity

With Prompt- Based User Interface like ChatGPT or BARD, Although User with Natural Language Interaction can perform multiple tasks without the need for specialized programming or training with Few-Shot Learning (Instruction with Example) and Zero-Shot Learning (Instruction without Example). Yes, it could be a handful of replacements for very complex UIs. But it all depends on how much flexibility and rigidity we need to provide for users with considering AI limitations.

Writing prompts to get a desired output from AI is not a skill everyone has. I know many people in my personal circle who struggles to write a simple leave letter and resume in English. Not because they are illiterate, but because they are just happy with delegating that work to others and working with templates. SourceAlso, a recent article by NN Group says about Prompt Driven UI“major usability downside is that users must be highly articulate to write the required prose text for the prompts. According to the latest literacy research (detailed below), half of the population in rich countries like the United States and Germany are classified as low-literacy users. (While the situation is better in Japan and possibly some other Asian countries, it’s much worse in mid-income countries and probably terrible in developing countries.)” 2. Probabilistic Nature of AI will demand new UX solutions when w incorporate AI technology in products.

The Nature of the Large Language Model (AI) technology behind ChatGPT and BARD-like products is purely probabilistic. In simple terms, If the user sends input, it smartly guesses the output with respect to the input based on the memory/data it already has. This approach might be not offensive for creative tasks but definitely it offends for noncreative tasks and business models running around it.

SourceDue to this nature, the probability of getting undesired results or partially desired results is high than getting desired results. So the user needs features to manage these probabilistic natures of AI. Regenerate Response in ChatGPT is one initial feature that falls into this category.

The worst part is If the user is not capable enough to assess the output. then wrong output might mislead the user and might affect the user's credibility and reputation. Which will overall lead to a bitter User Experience. This is where Responsible AI solutions are highly needed.

Conclusion: Understanding this AI Goldrush as a UX professional

Venture Capitalists and Business Enterprises will be always interested in exploring the potential next big thing Because none of the above people want to become victims of disruptive technology and also they don't want to lose the opportunity cost. Two years before most of us thought BlockChain and VR will be the immediate next big thing, But LLM(AI) came as a surprise to Non-AI Enthusiasts. Honestly, only after certain years of experimentation by multiple entities, we all will really know the real impact of AI in Industry. Whatever happened after the ChatGPT release shows the feasibility and access of AI technology are comparatively better than any other promising next-big things. So enterprises will obviously continuously try to explore, adopt the optimal ones, and market this technology with or without its actual need for it. So most of our business stakeholders expectation from UX professionals ‘to deliver better User Experience design in AI products’ And they will really care less about ‘Using AI in UX Design Process” until the entire business runs around it.

Feel Free to connect with me on Linkedin! https://www.linkedin.com/in/isanthoshgandhi

Sources & References

1. Prompt-Driven AI UX Hurts Usability (https://www.linkedin.com/pulse/prompt-driven-ai-ux-hurts-usability-jakob-nielsen/)

2. Whose job does AI automate? (https://ie.linkedin.com/posts/kozyrkov_whose-job-does-ai-automate-activity-7060264121262338049-rRfh)

3. How ChatGPT/BARD/LLM Will Change UX/HCI Design Forever?(https://www.linkedin.com/pulse/how-chatgptbardllm-change-uxhci-design-foreverand-save-mauro-chfp)

4. Large Language Models for Product Managers: 5 Things to Know(https://www.assemblyai.com/blog/large-language-models-for-product-managers-5-things-to-know/)