‹ Back to Video Series: AI UX

AI UX: Minimize Intrusion

  • Overview
  • Resources
  • Transcript

This episode of AI UX looks at how being thoughtful with your data gathering and adopting a less is more approach can benefit you and your users.

AI UX YouTube* Playlist

Subscribe to the YouTube* Channel for Intel® Software

Additional Resources:

Loi, D., 2018. Intelligent, Affective Systems: People’s Perspective & Implications. Proceedings of CHIuXiD 2018, Yogyakarta, Jakarta, Malang, Indonesia.

Loi, D., Raffa, G., and A Arslan Esme, 2017. Design for Affective Intelligence. 7th Affective Computing and Intelligent Interaction Conference, San Antonio, TX.

Bostrom, N., and Yudkowsky, E., 2014. The Ethics of Artificial Intelligence. The Cambridge Handbook of Artificial Intelligence, Cambridge University Press.

Sophia Chen. AI Research Is in Desperate Need of an Ethical Watchdog. Accessed October 14, 2017.

PwC. 2017. Global Artificial Intelligence Study: Sizing the Prize.

Gershgorn, D. 2017. The Age of AI Surveillance is Here. Quartz

Harari, Y.N., June 23, 2017. Are We About to Witness the Most Unequal Societies in History? The Guardian.

This is AI UX, a miniseries focused on ten guidelines created to assist you in the design and development of AI-based systems. I'm Daria Loi, an Intel researcher. And today I will talk about guideline number two: adopting the minimize intrusion mantra with a less is more data approach.

The more data you fit into a system, the more the system will learn, improving its algorithm's efficiency and accuracy. Having said that, you must be thoughtful in your data approach. When designing and developing AI systems, use minimalism as a compass. This means to only collect the data that is necessary for you to successfully achieve the requested transaction.

For instance, if I recorded you 24/7, the likelihood for me to accurately understand your life and use algorithms to predict your future behaviors would be higher than if I only had data from a handful of motion sensors in your home. Yet 24/7 monitoring would be intrusive and exhausting. You would probably be hesitant to let someone collect data on you all the time, as your personal privacy would be compromised.

Just because a massive data collection could make your AI-based system more advanced, it does not mean that it should be pursued fully. There has to be a level of respect and anonymity for your users to ensure that people will embrace and use an intelligent system. Research interviews I conducted made it clear that the more someone becomes familiar with an intelligent system, the more they trust and feel comfortable in using it.

To establish familiarity, careful design and considerations are needed. One of the participants in my research was originally skeptical of smart home systems, explained that she considered the technology only after her husband purchased the system.

Once the system reached their home, she took some time to get familiar with it. As she grew more familiar and her comfort with the system increased, she started exploring advanced capabilities and felt open to embrace more systems to refine her smart home abilities, even if that meant allowing her systems to collect more data. She was able to do this after realizing the value of the technology through daily use, validating that it was a safe technology, and after realizing that the system value-add justified additional data collection.

Familiarity and trust are crucial if you want someone to feel confident using your AI system. This process takes time, patience, and respect for user preferences.

Based on these findings, I have concluded this recommendation. Adopt the minimize intrusion mantra and a less is more data approach.

The minimize intrusion mantra is something you should always be saying to yourself while designing for user experience. You can follow this mantra by setting the system default settings and a basic level with basic function.

Additionally, the less is more data approach should be followed by collecting, using, and storing only the necessary amount of data. Make it so the user can change these things later on once familiarity is established. A smart system should dynamically change settings based on direct user's request or feedback loops. Keeping your design less intrusive and being thoughtful about your data collection and usage will greatly improve your user experience design and will help the user be comfortable enough to try your experiences again.

Thanks for watching. Don't forget to like this video and subscribe. And I will see you next week on Tuesday for more AI UX.