Artificial Intelligence (AI) refers to computer systems that can perform tasks that typically require human intelligence. These tasks include learning, reasoning, problem-solving, understanding natural language, and adapting to new information.
AI applications range from virtual assistants and image recognition to complex tasks such as autonomous vehicles and medical diagnosis. They contribute to sustainable design by optimizing resource use and reducing waste.
In this video, AI Product Designer Ioana Teleanu talks about how AI is changing the world.
AI systems use algorithms and computational models to analyze vast datasets, identify patterns, and make decisions. Machine learning, a subset of AI, enables systems to improve performance over time by learning from experience without explicit programming, which can support a circular economy by enhancing resource efficiency and recycling processes. Deep learning, a specialized subset of machine learning, centers around deep neural networks with multiple layers, which mimics the human brain's complexity. These networks autonomously extract intricate patterns from extensive datasets, enabling advanced capabilities like image recognition and natural language processing, and can be applied in incremental modular design to iteratively improve products and systems.
© Interaction Design Foundation, CC BY-SA 4.0
Artificial Intelligence encompasses a spectrum of capabilities, from specialized task-oriented systems to intelligence that mirrors human cognitive functions. At the core of this distinction lies the difference between Narrow AI, also known as Weak AI, and General AI, also known as Strong AI or Artificial General Intelligence (AGI).
Narrow AI refers to systems tailored for specific, well-defined tasks within a limited scope. Examples of narrow AI models are common in our daily lives, from voice recognition tools like Siri or Alexa to recommendation algorithms powering platforms like Netflix and Spotify. Chatbots assisting with customer service on websites and specialized image recognition software in facial recognition or medical imaging analysis are also instances of narrow AI. Its defining characteristic is its lack of capacity to generalize knowledge beyond its designated domain.
On the other end of the spectrum is General AI, an advanced form capable of comprehending, learning, and applying knowledge across various tasks—mimicking the breadth of human intelligence. Unlike narrow AI, AGI can reason, problem-solve, adapt, and exhibit self-awareness. The ultimate goal of AGI is to perform any intellectual task that humans can, seamlessly transfer knowledge between domains, and autonomously improve over time.
While narrow AI excels in specific functions, AGI is the pinnacle of AI development. Currently, however, most AI systems are narrow, designed for specialized tasks and lacking the broad adaptability of AGI. Achieving AGI remains the significant and ambitious objective of AI research and development.
© Interaction Design Foundation, CC BY-SA 4.0
On the spectrum of AI, generative AI is positioned between narrow and general AI. It’s a category of artificial intelligence that focuses on creating new content, data, or artifacts rather than performing specific predefined tasks, embodying the principles of humanity-centered design to ensure ethical and societal benefits. It involves machines that can produce outputs, such as images, text, or other forms of content, that weren't explicitly programmed into them. Generative AI often employs deep learning and neural networks to learn patterns from large datasets to generate novel outputs. Outputs are created in response to AI prompts. Effective prompts, or prompt engineering, are an essential part of human-ai interaction.
ChatGPT, a generative language model by OpenAI, was released in 2022. Within five days, over a million people had signed up for it. Unlike traditional programs with fixed responses, ChatGPT can dynamically generate answers based on the patterns it learned from vast amounts of text data. This ability makes it versatile—you can ask it questions, request information, or even use it for creative writing. This type of AI is valuable for various tasks, from aiding in research to helping with creative projects.
DALL-E, another application from open AI, generates images. Similar to ChatGPT, it creates a unique output from text inputs or prompts. For instance, you can ask DALL-E to generate an image of a "giant rubber duck" or a "surreal cityscape with floating buildings," and it will produce an original image matching that description. This kind of AI is part of the broader category of generative models designed to create new content. DALL-E showcases how AI can be used for artistic and creative endeavors, offering users a new way to generate visual content.
AI-generated art refers to artworks that are created with the assistance or direct involvement of artificial intelligence. In this process, artists or an individual collaborate with AI systems, which can include machine learning models and generative algorithms. These AI tools analyze vast datasets and learn patterns to generate new artistic outputs. AI-generated art spans various forms, including visual arts, music, literature, and more. The unique aspect of AI-generated art lies in the fusion of human creativity with the computational capabilities of AI, challenging traditional ideas of the arts and opening up new possibilities for artistic expression.
Unsupervised from Refik Anadol's Machine Hallucinations project, is a fascinating example of AI-generated art. It exemplifies the intersection of technology and creativity. Unsupervised, a product of deep learning algorithms processing vast datasets from the Museum of Modern Art (MoMA), generates abstract images guided by intricate patterns and associations within the museum's collection. This artwork is a testament to the capabilities of generative AI—its potential to create unique and unexpected outputs beyond explicit programming.
Learn more about AI-generated art, its challenges and opportunities in this video.
Transparency becomes a crucial concern as the origin of information and the decision-making process of these AI systems can be obscured. The potential for bias, privacy implications, and the need for explainability in AI-generated content underscore the intricate landscape that artists and technologists navigate.
In this video, UX design pioneer Don Norman, talks about how we can collaborate with AI—AGI is not a reality just yet so the AI apps we use need human input.
UX design pioneer Don Norman warns that these programs are not truly intelligent yet. They don't have wants, needs or a sense of self as humans do. Instead, they make decisions based on patterns in data too large for humans to process.
AI follows a complex set of logical rules called algorithms. Multiple algorithms connect in a way that mimics the human brain, called a neural network. This network can learn and improve its process over time. We call this "machine learning."
Artificial intelligence has already improved technologies like voice recognition and language translation. Even still, AI has shown even more potential and some surprising new applications.
For example, AI can create art and literature in the style of human authors and artists. Yet, they don't express emotions or create their own artistic style without human help.
This emerging technology has a variety of exciting and frightening uses. AI programs make it easy to pretend to be someone else or pass off AI content as your own. On top of that, the ethics of sentient AI will be a hot topic in various fields as the technology advances.
Text Generators
ChatGPT: This program can write new text by comparing itself to similar works on the subject.
Bard: A chatbot by Google used to create a more intelligent and conversational search algorithm.
Image Generators
Midjourney: An image generator that uses subject and style prompts to create new works of art.
Dalle-2: Similar to Midjourney but specializes in realistic images.
Video and Speech
Gen-1 Video editor: A video editor that shifts a video into a different style. For example, making a live-action video into an animation.
DeepFaceLab: One of many programs that make "deep fakes." Deep-fakes are videos that change faces and voices to impersonate other people. The most famous example is Jordan Peele’s Obama deep fake video from 2018.
Dragon Speech recognition: This program learns speech patterns to turn speech into text. It was the basis of most modern speech recognition software.
UX Design
Galileo AI: Entire user interfaces can be generated based on text prompts.
Genius: The AI design companion for Figma that fleshes out a full layout from a few design elements.
Interaction designers use AI technologies in a variety of ways. Artificial intelligence improves search algorithms for web searches, streaming services and other platforms. They can analyze terabytes of data to find patterns a human brain couldn't.
There is no doubt that AI will change how users interact with products and services. AI voice assistants and chatbots are examples of interfaces that adapt to user inputs in real-time. UX designers design the voice and the functions of voice assistants to appeal to users. Even though chatbots are text, they still need to make sense in the product's context of use. Like any interface, designers want to make a user experience that users trust and enjoy using.
“There’s a very simple formula, perceived trustworthiness plus perceived expertise will lead to perceived credibility.
Since AI is in service to human beings, I can't imagine a case where UX isn’t relevant…If you blow the UX design, it doesn't matter how good the AI is.”
-Dan Rosenberg, UX Professor at San Jose University.
The goal of artificial intelligence today is to be credible. They should be reliable tools and assistants for humans performing specific tasks. This credibility comes partly from a well-designed user experience and intuitive user interface.
The potential for AI to replace human workers is possible. But, it is more likely to be used to assist humans in making decisions. For example, AI could assist in usability tests or find patterns in user feedback or other user research tasks. AI has the potential to transform the essential tasks of a UX researcher.
“When it comes to [user] research, it is such a strategic discipline I can't think that we will ever automate it. If we are talking about general usability testing, that is going to be something where AI is going to play a big role. AI does something extremely well and that’s pattern recognition.”
-Greg Nudelman, Head of Design at LogicMonitor and Author on UX for AI
Many experts see the potential for AI to change human-computer interaction but also have doubts. AI systems can improve data analysis, assist translation, and help creatives bring ideas to life.
Yet, all this brings up deep ethical questions. Creatives of all types are forced to compete with AI, which can plagiarize their work in minutes. The question of who owns that AI content is also unclear.
Some communities have banned AI art entirely, even as the ability to tell them apart from human work diminishes. Even if AI does not fully replace humans, what will our economy or workplace look like if AI replaces daily tasks or even jobs?
In the future, “Strong AI” would learn, think, and generally function on the same level as humans. As fully sentient beings, there are moral questions of ownership and legal definitions of autonomy to grapple with.
Despite these challenges, tech companies are investing heavily in AI to explore the possibilities.
Discover how to design for AI and how you can incorporate AI tools into your design process in our course, AI for Designers.
For more on the role of AI and other technologies on design, take our course: Design for a Better World with Don Norman.
Norman, Donald A. Design for a Better World: Meaningful, Sustainable, Humanity Centered. Cambridge, MA, MA: The MIT Press, 2023.
Read more articles and essays by Don Norman on JND.org.
Watch our Master Class: AI-Powered UX Design: How to Elevate Your UX Career with Ioana Teleanu.
Watch our Master Class: How To Design for and With Artificial Intelligence with Dan Rosenberg.
Watch our Master Class: How To Design Experiences for AI with Greg Nudelman.
To learn more about the differences between AI, machine learning, deep learning and neural networks, read AI vs. Machine Learning vs. Deep Learning vs. Neural Networks: What’s the difference?.
Read more about AI and its various applications in the Investopedia article, Artificial Intelligence: What It Is and How It Is Used.
Everything About Usability Testing Through Al-Powered Software
Do you want to improve your UX / UI Design skills? Join us now
You earned your gift with a perfect score! Let us send it to you.
We’ve emailed your gift to name@email.com.
Do you want to improve your UX / UI Design skills? Join us now
Here’s the entire UX literature on Artificial Intelligence (AI) by the Interaction Design Foundation, collated in one place:
We believe in Open Access and the democratization of knowledge. Unfortunately, world-class educational materials such as this page are normally hidden behind paywalls or in expensive textbooks.
If you want this to change, , link to us, or join us to help us democratize design knowledge!
