5 Ways AI Will Affect Design In The Future

5 Ways AI Will Affect Design In The Future
Design

David Fisher, senior product designer at digital product studio ustwo, discusses how AI can be leveraged as a tool for design problem solving

PSFK Op-Eds
  • 3 july 2017

AI’s recent resurgence has garnered considerable media attention and has prompted many debates – particularly about how this transformative technology will change the workplace. One of the biggest questions at the moment in the creative industry is “how will AI affect design?” Many notable players in the industry have offered high-level viewpoints and have even created initial resources; yet, there is little in the way of practical advice for designers. The guide below offers five considerations for designers wanting to get a foothold with this technology and how it can be leveraged as a tool for problem solving.

  1. AI approaches through the ages
    Over the preceding 5 decades, many technologies have been created in the development of the field of AI. It is important to catalogue the difference between these, as there are differences in capability, implementation and performance when used to create products and services.

Early AI

Early AI systems arose from an approach where a series of logical rules were used to model intelligence. These systems had practical limitations and tended to work best in situations where queries were simple, and the scope of the subject was narrow. These systems exist in many formats today, including basic chatbots and expert systems that inform decision processes based on a limited series of inputs.

Contemporary AI systems

Contemporary AI is fundamentally different from early AI. New approaches leverage the use of large datasets on which algorithms can be ‘trained’ – essentially, identifying patterns in the data that then enable reasoning and decision-making. There are many subsets of machine learning and different methods by which each system can be trained. A comprehensive resource for learning about the history and development of AI is Russel & Norvig’s AI: A modern Approach.

  1. High quality data is the bedrock of AI  

A challenge many teams will face when building AI products and services will be securing data and ensuring that it is in a usable format and complete enough to solve a business problem.

Data cleanliness and completeness is important because it impacts how well a model will be able to differentiate between patterns in the data versus noise. Data scientists are well poised to guide teams through this process; however, in the absence of such guidance, Patrick Hebron’s “Machine Learning for Designers” provides best practices for teams looking to sanitize their datasets.

A secondary problem teams might face when gathering data is related to the way organizations are set up. Large companies are often segregated into hierarchies or independent business silos, each with discrete marketing, product and engineering teams. This means that data collection may not be a well-defined or standardized process within various parts of the organization. Teams in these situations will be well served by planning ahead and putting together business cases and presenting ideas of the capabilities of potential AI products and services to win the support of stakeholders.

  1. Models, Models, Models

If you’ve been following AI in any capacity, chances are you will have heard the term AI model. In very simple terms, an AI model can be considered a system’s ‘experience’. Essentially every item of data that the model is trained on gives the system a reference point to make predictions when presented with a new input.  A compelling example of how this works in practice can be seen by experimenting with ClarifAI’s various pre-trained models, which range from face detection to identifying ingredients in images of food and offer a shallow learning curve for those getting started.

There are presently many proprietary offerings on the market, open source as well as MLaaS (machine learning as a service). Teams are well advised to weigh the options when it comes to using off-the-shelf solutions or building out their own models and infrastructure. The latter represents considerably more complexity over the former, particularly for common tasks such as image classification.

  1. New interaction paradigms

The success of Amazon’s Echo is a great example of how advances in speech recognition and natural language processing create opportunities for natural and expressive interactions with products and services. By moving beyond the graphical user interface, designers can now begin to explore alternative methods of interaction enabled by AI capabilities.

This shift in interaction paradigm presents a challenge for designers. New interactions should fulfill a user need, while being sensitive to context. Performing thorough research and testing in diverse situations can ensure designers create interactions that are valuable and contextually appropriate.

  1. With great power comes great responsibility

As AI begins to seep into products and services, things will inevitably wrong. This poses a new consideration for designing AI-powered systems – what is the procedure when a system makes a mistake? Examples of this happening are already plentiful, with one of the earliest examples occurring in 2015 when Google Photos’ face-recognition model labeled a user’s images in the worst imaginable way.

While Google has since rectified the issue, designers working with AI would be prudent to familiarize themselves with best practice principles to reduce the likelihood of things going awry.

Designers can mitigate risk by learning more about how datasets can be biased, as well as common examples of how systems can be fooled. Designers can also design systems, which ensure end users retain some level of control by prioritizing immediate feedback on any erroneous behavior. Even basic measures can minimize the potential fallout if something goes wrong and ensure that the quality of the end user experience can be preserved and improved over time.

If designers can adhere to these five considerations when it comes to AI, they’ll not only give themselves an edge in the most pertinent area that’s mystifying the creative workplace, but they’ll also prime themselves for a future that will allow them to do their best work to date.

David Fisher, senior product designer at digital product studio ustwo, heads up innovation strategy and product development, along with the coordination of internal and external business/design strategies for the company.

AI’s recent resurgence has garnered considerable media attention and has prompted many debates – particularly about how this transformative technology will change the workplace. One of the biggest questions at the moment in the creative industry is “how will AI affect design?” Many notable players in the industry have offered high-level viewpoints and have even created initial resources; yet, there is little in the way of practical advice for designers. The guide below offers five considerations for designers wanting to get a foothold with this technology and how it can be leveraged as a tool for problem solving.

+AI
+AI
+Amazon Echo
+artificial intelligence
+chat bot
+data
+Design
+machine learning
+Public
+technology

Learn About Our Membership Services

Need Research Help?
As a member you can ask us any research questions and get complimentary research assistance with a 4-day turnaround. Reports inclde stats, quotes, and best-inclass examples on research topics.
Remain Informed & Strategic
We publish several trends reports each month. By becoming a member you will have access to over 100 existing reports, plus a growing catalog of deep topical analysis and debrief-style reports so you always remain in the know.
See Trends Come To Life
Meet your peers and immerse yourself in monthly trend and innovation webinars and discounted conferences.
No search results found.