Skip to content Skip to sidebar Skip to footer

As Machine Learning and AI Perform Magic, How Can UX Professionals Help?

The self-driving car with no human driver behind the wheel is no longer a futuristic fantasy. According to recent business articles including Jonathan Camhi’s article about Ford in December 2017’s Business Insider and Samuel Gibbs’s November 2017 article about Waymo in The Guardian, this autonomous revolution will not only disrupt jobs and economies, but also will change the way people commute and transport goods. Per Simon Lewis in Time magazine (June 2016), Carolin Cummins in the Sydney Morning Herald (October 2016) and a Reuters article (April 2015), robots are the new frontline receptionists in some hospitals, shopping malls, and offices to greet visitors, provide directions, and even do some administrative work such as check-in for meetings and contacting hosts.

Artificial intelligence (AI) is a collection of technologies that enables machines to perform tasks that previously would require human intelligence, such as learning, reasoning, solving problems, and adjusting to new input. The technology can augment and even replace human activities. Recent research on the impact of AI, described in the Accenture Technology Vision (2017), reveals that AI could double annual economic growth rates in 2035 by changing the nature of work and creating a new relationship between man and machine. The transformational effects of AI in the coming decade will impact businesses in manufacturing, retail, transportation, agriculture, finance, healthcare, law, advertising, insurance, entertainment, education, and virtually every other industry, per Erik Brynjolfsson and Andrew McAfee’s cover article in the Harvard Business Review (July 2017), and Dom Galeon’s article in the webzine on the first nation with a state minister for AI (December 2017).

Figure 1: Human/Robot

As user experience (UX) designers and researchers, we have the responsibility to make these intelligent solutions trustworthy, friendly, and useful, avoiding bias as far as possible. Most importantly, we must make sure the technology helps instead of harms humanity and society, as predicted in multiple articles, including Steve Levine’s summary of a McKinsey study that predicts 1/3 of America’s workforce may be wiped out by 2030 (November 2017).

At the UXPA 2017 International Conference, a collaborative workshop was held in which 21 UXPA attendees, including UX researchers, designers, and consultants, came together to discuss new challenges and opportunities in the new AI and Machine Learning (ML) age. Specifically, we asked the following questions:

  • How will our roles change?
  • What new skills should we acquire?
  • What should we do to help to make positive impact?

How Will Our Roles Change?

With smart systems it is more important to understand the roles of human and machine. Is the machine a tool, an agent, or a replacement for the human currently performing the job? Before we design any systems, we must understand the needs of the people, their environment, and how technology could best support them. We should work together within our teams to establish a shared vision for the user experience.

We have the responsibility to use our knowledge of human and social behaviors to help all team members, including decision makers and customers, to understand the potential consequences of the technology. For these types of systems, we also need to consider things like training data and potential biases, changing experiences over time, and systemic implications of a technology that learns and changes. As researchers, we should apply appropriate research methods to help the team understand the current user journeys and experiences. Based on this, and using our understanding of human capabilities, limitations, and behaviors, we can help envision the future experience and identify possible consequences. With our help, the project team can think about the human-machine ecosystems holistically, ensuring that human factors like complex behaviors, emotions, diversity, culture, and norms are considered, establishing appropriate feedback loops and determining how to measure impact. And finally, we need to use the right testing methods to help validate our assumptions.

Smart systems demand expanded user scenarios, including more actors and different kinds of impacts:

  • One example is Roomba, the automatic vacuum, where researchers and designers need to consider scenarios from learning, setting up, and running the device safely, to maintaining it. In addition, we may need to consider the excitement of the family who purchased it and how the children will treat it.
  • Another example is Siri and other virtual assistants, where the wrong tone (“How can I help you?”) may lead users to assume a broader set of abilities than are possible (“Siri, which car should I buy?”).

As designers, we not only continue to consider the information architecture, interaction, and visual designs, we also need to consider the experiences of setting up, tuning, and dealing with interruptions, connections with other systems, malfunctions, and so on.

UCD methods in a workshop
Figure 2: UCD methods in a workshop

What New Skills Should We Acquire?

Some aspects of our UX work may disappear or significantly change. For example, some workshop attendees reported their company applied AI to generate customized graphic banners during a hot shopping season, which augmented the work of graphic designers and helped to meet the demand. AI can also automate or augment some research and analysis tasks to help us better understand people’s needs and behaviors. Whether we like it or not, the train has started, and we must grow and expand our scope to stay current.

First, we need to have a deep understanding of AI and ML, how they evolve, and what the technology changes will mean to humans and society. We need to have a good understanding of the technology landscape and the ecosystem which consists of machine, human, and the array of data that shapes the experience. We may need to acquire new knowledge about data science and programming to better work with the professionals in those areas. We need to understand what data will feed the training system and how to identify if the system is biased.

User research and result findings will play bigger roles in defining the technology. User researchers and designers need to learn how to better communicate with—and influence—decision makers, and collaborate with new types of stakeholders such as AI research scientists, robotics scientists, or neural networks/machine learning specialists.

We also need to learn how the feedback loop works in the technology and how to gather feedback from the user to help the intelligent system learn, present changes to users, and measure the impacts.

What Should we do to Make Impact?

1.    Learn and share

Workshop participants suggest UXPA create an infrastructure for group learning and collaboration. For example:

  • Establish a platform for collaboration
  • Improve our strategic thinking and soft skills to help us better engage decision makers and executives in a way that may not previously have been done
  • Enlist interested members to share valuable content and annotate it for the UXPA community
  • Identify areas for further research related to behaviors, experiences, and broader impact
  • Invite AI and data science professionals to work with us and create a shared understanding of the technology and its impact on humanity
  • Identify suitable books and online courses to help us improve both hard and soft skills

2. Develop guidelines and best practices

Develop guidelines and best practices to help practitioners play more strategic roles in the organization, ensuring the technology brings positive impact to both people and their environment. Develop guidelines, standards, and best practices for new interactions, such as conversational UIs, ambient intelligence, augmented and virtual reality, and the like..

We must describe processes that will define the full user experience lifecycle from initial use and learning of the product, system, or set of technologies, to regular usage, including what users can do when it does not perform as expected. Also, we need to consider technologies that are not explicit or user-facing but have an impact on users. We must consider the disruptive nature of the technology, its impact on existing experience, and the environment that surrounds the user.

We should also develop guidelines and materials to help UX professionals educate the public on how smart tools and systems may affect people’s behaviors and feelings.

3. Make an Impact to Help Humanity

Workshop participants suggested UXPA members work with UXPA to craft a standard message about humanizing AI. For example, the technology must consider the implications of ethics, trust, and privacy. Some participants suggested UXPA form a task force for this effort.

Workshop presentations
Figure 3: Workshop Presentations

Looking toward the future, UX practitioners can help define governance that works to ensure ethics and privacy and takes a strategic role in managing the impact of unintended consequences. This topic was discussed and elicited  more questions than answers:

  • As there are more and more smart systems, help imagine and humanize the future. For example, as people become addicted to their systems, as we are starting to see with some games and social media platforms, how can we bring them back to the real world to seek more human contact?
  • Can we find ways to bring the issues to the broader public for discussion rather than addressing issues within the current silos of corporations and governments?
  • One company has already sent UX representatives to speak at the United Nations. This kind of national and international policy cooperation is needed soon. Can UXPA be the organization to spearhead this kind of work going forward? If not, then who should do it?


Workshop participants agree that UX designers and researchers need to be the co-creators of intelligent solutions to make sure AI technology works for people and society. More than ever, we must consider the capabilities and roles of human versus machine. When should machines make decisions and take action, and when should they augment or support people making decisions? How will these AI solutions make people feel? Do people feel like the solution is trustable, easy, and fun, or do they feel frustrated or even potentially endangered?

UX professionals must act to learn, share, collaborate, and participate in cognitive technology research and development both at a strategic level and as a part of the product development process. We should also get involved in governance. We encourage UX professionals to join us and continue this dialog so that we can help create a better world.


We want to thank all the workshop participants at the 2017 UXPA International Conference for contributing the ideas outlined in this article. Special thanks to Alasdair Stuart-Bell, Brenda J. Burkhart, Chang Ting “CT” Yu, Jeff Johnson, and Bernardo Doré, who submitted position papers before the workshop to help us prepare for the discussion topics. Thanks also goes to Karen Bachmann, Duane Degler, Brenda J. Burkhart, and Jeff Johnson for reviewing the article and providing suggestions. Finally, we want to thank the student volunteer, Kristine Malden, for helping with the workshop logistics onsite.