The Brief: “do some quick research into how AI is being used within your own specialism”.
I wanted to understand how Artificial Intelligence (AI) is currently used to enhance the User Experience (UX). I found a very informative online Stamford Lecture that I have tried to summarise in this post with some additional information from other sources.
AI can be used to augment the user experience using several different techniques. Research in to this field has shown that the effectiveness is dependent on the certainty of response:
Automation: When the certainty is extremely high the system will automatically execute the action. For example, if you write an attachment in the header of an email but have not attached anything you will get a pop up asking if you meant to attach something. (Bernstein, 2020)
Adaptive Interaction: This is where the interface is tweaked to bring things closer to the user rather than execute something automatically. The interface aids by suggesting things that you might need promptly. For example, the tools used most in an application are readily available and easy to access and the other content is collapsed. Some AI can be utilized to detect the type of device being used so if it detects a touch interaction it can change from the mouse function to touch screen and resize content to fit the device (responsive). (Bernstein, 2020)
Accelerators: One version of an accelerator is the auto-complete function, for example, if you started typing in the Google search engine before you finish typing Google may come up with a list of suggestions that you can click on.
If you are using a map application the system may suggest a faster route. In this case, the user will have usually have control to opt-in or continue with their original course. (Bernstein, 2020)
Suggestion: Suggestion can be used when there is lower certainty, the system will provide a menu of personalised options for example Netflix recommendations. With a large number of personalised suggestions, there is a higher chance that one of the suggestions will be fit for the end-user. (Bernstein, 2020)
Do Nothing: When it comes to low certainty the best approach is to give the user full control with Direct Manipulation. (Bernstein, 2020)
Another way that AI has been utilised to enhance UX is through Smart Chatbots. Chatbots can answer customer quires in real-time without the need for human intervention. There are two types of chatbots one is rule-based and will complete actions following a script, the other uses AI and Machine Learning (ML). The AI-driven Chatbot can create a more natural closer to human conversation. This is a great solution for some quires however if the Chatbot can not provide the information the customer requires they will be handed over to a human agent.
AI with Biometrics
AI has also been paired with biometrics to improve the customer journey when it comes to call handling. For example, a user will use a phrase to get a call directed to the right department rather than having to go through numerous menus and selections. It can also be used for security with voice recognition so the user doesn’t have to answer as many security questions to speak to an advisor or access information.
AI with other devices
AI becomes more powerful and more necessary when it is moved away from the desktop to other Smart technologies. For example, smart speakers, “Smart speakers rely on a set of complex artificial intelligence (AI) technologies. They listen to sound waves and convert them into words using automatic speech recognition (ASR). Then, they convert those words into meanings using Natural Language Understanding (NLU). Once the meanings are understood, the smart speaker responds using Natural Language Generation (NLG) technology. These relatively small devices mask layers of complex technologies to deliver a simple, frictionless user experience.”(Elnaj, 2019)
Limitations of AI
When using AI as UX designers we need to be cognizant of how we approach interaction design to manage uncertainty and deliver a good user experience. AI is a very powerful tool but it has its limitations:
- Models can inherit bais’ (gender/race etc.) this can come from both the algorithm and the data.
- IA requires large amounts of data from the training through to execution. If you want to implement machine learning the data set should be chosen with the goal of enabling learning.
- AI lacks emotional intelligence – “cannot classify human sentiments and mindsets into one-of-a-kind data points or profiles”
- There will also be some instances or use cases where the user will only feel happy with human interaction.
- There can be gaps between the user intention and machine response. (Dialani, 2019)
In conclusion, I think that AI can improve the customer experience when the users are considered in the design. I see AI as being complementary to UX. I think that with its current limitations there is still a clear role for a human UX designer. I believe that AI will help advance user research by supporting data collection and analysis.
Dialani, P., 2019. Top 5 Limitations of Artificial Intelligence | Analytics Insight. [online] Analytics Insight. Available at: <https://www.analyticsinsight.net/top-5-limitations-artificial-intelligence/> [Accessed 8 February 2021].
Elnaj, S., 2019. Council Post: The Rise Of AI-Enabled Smart Speakers And Their Future In Our Lives. [online] Forbes. Available at: <https://www.forbes.com/sites/forbestechcouncil/2019/07/08/the-rise-of-ai-enabled-smart-speakers-and-their-future-in-our-lives/?sh=3c53ae0f35ab> [Accessed 8 February 2021].
WhosOn. n.d. Everything you need to know about machine learning chatbots. [online] Available at: <https://www.whoson.com/chatbots-ai/chatbot-learning-everything-need-know-machine-learning-chatbots/#:~:text=Introducing%20machine%20learning%20chatbots,between%20computers%20and%20human%20language.> [Accessed 8 February 2021]
Six, J., 2019. Artificial Intelligence and User Experience. [Blog] UX Matters, Available at: <https://www.uxmatters.com/mt/archives/2019/11/artificial-intelligence-and-user-experience.php> [Accessed 8 February 2021].
Fitzgerald, A., 2020. How UX professionals can plan for AI as it shapes the future of the industry | UserZoom. [online] Userzoom.com. Available at: <https://www.userzoom.com/ux-library/how-can-ux-professionals-plan-for-ai/> [Accessed 8 February 2021].
Bernstein M, Learn.stanford.edu. 2020. Free Webinar: Create a better user experience. [online] Available at: <https://learn.stanford.edu/ai-for-ux-wbn-on-demand.html> [Accessed 8 February 2021].
Nielsen, J., 2016. AI’s Impact on New Technologies (Video). [online] Nielsen Norman Group. Available at: <https://www.nngroup.com/videos/ais-impact-new-technologies/> [Accessed 8 February 2021].
Taulli, T., 2019. Artificial Intelligence (AI): What About The User Experience?. [online] Forbes. Available at: <https://www.forbes.com/sites/tomtaulli/2019/04/27/artificial-intelligence-ai-what-about-the-user-experience/?sh=4b1ee48516fe> [Accessed 8 February 2021].