A Beginner’s Guide To Designing Trustworthy AI
Why are we still using smart assistants as music players? In medical diagnosis, where the stakes are really high, would you trust an AI system to detect your disease based just on some past data that it gathers?
Rapidly evolving AI technology has just started to show its cards — that can be used for good( like inclusive design) or not so noble intentions (like unsolicited, unregulated and targeted advertisements) which might lead to consequences satisfactory or undesirable. It is our responsibility as creators of products to make the experience of our users less terrifying when they are entrusting their information with us to make their lives better.
If you are new to the field of building AI products and are planning to build one, this is just the right repository of resources to get started-
What Makes An AI Product Trustworthy?
The following resources give a sneak-peak into the principles that govern the design of a trustworthy AI:
Microsoft’s approach to building trustworthy AIs
Summary: Not having an understanding or being in a “black box” of how the AI product came to a particular decision can be baffling and reasonably untrustworthy. Hence Transparency is one of the 6 principles of Microsoft to be designing AI to be trustworthy, the others being: Fairness, Inclusiveness, Reliability & Safety, Transparency, Privacy & Security and Accountability. Also check out Page 57 onwards in the book — The Future Computed, where it has been explained what these principles mean with examples. This resource can be a great starting point to analyze the trustworthiness of an AI system.
UX for AI: Building Trust as a Design Challenge
Summary: The author takes TED talk by British philosopher Onora O’Neil as a reference who points to three components of trust: competence, honesty and reliability and uses it as reference and analyses trustworthiness of AI in different products taking these components of trust as perspective.
AI UX: 7 Principles of Designing Good AI Products
Summary: This article gives really good principles one must consider while designing AI products like: “We should let people know if an algorithm has generated a piece of content so they can decide for themselves whether to trust it or not”…”we should give users hints about what the algorithm does or what data it uses.”…
Responsible bots: 10 guidelines for developers of conversational AI
Summary: 1. Articulate the purpose of your bot and take special care if your bot will support consequential use cases. 2. Be transparent about the fact that you use bots as part of your product or service. 3. Ensure a seamless hand-off to a human where the human-bot exchange leads to interactions that exceed the bot’s competence. 4. Design your bot so that it respects relevant cultural norms and guards against misuse. 5. Ensure your bot is reliable. 6. Ensure your bot treats people fairly. 7. Ensure your bot respects user privacy. 8. Ensure your bot handles data securely.9. Ensure your bot is accessible. 10. Accept responsibility.
How to build trustworthy AI products (Medium)
Summary: “Using AI is like any other technology that has tradeoffs… and you should be aware of the nuance. Build for the right amount of trust — not too much or not too little. Trust is based on performance, accountability, and transparency to humans.”
8 UX Surefire Ways to Design for Trust (Medium)
Summary: “Trust is the newest and strongest currency in our growing sharing economy and it is a key ingredient in building a successful product. It is more than being credible or reliable in the eyes of the user. When you launch a new product, what you are getting your customers to do is to take a leap of faith by getting them to use your product. How can we make that experience less terrifying and easy for the users? You start by designing for trust.”
Designing to Build Trust (UXMag.com)
Summary: “In this article, we will dive into the nature of trusted online experiences, why they are important, design attributes that we know people trust, and how design creates trust and distrust. We’ll illustrate the issues around designing for trust with a sample prototype of a healthcare exchange design and user reactions to it.”
How to Design for Trust in Digital Services (Prototypr.io)
Summary: “There is little doubt that trust is a crucial aspect as our society face rapid development in both social structure and digitalization. How could it be possible to find a universal language of trust? The answer to this is not a one-size-fits-all. We have to consider cultural differences, demography and the markets in which we operate.”
8 Tips for Building Trust in your Product (inc.com)
Summary: “In business, trust is earned and not given. When consumers trust your product, they’re happier and more engaged. User acquisition becomes easier. You put yourself in a position to make a real difference. However, trust can be elusive. It is tricky to build and easily lost.”
DARPA Is Working to Make AI More Trustworthy (Futurism.com)
Summary: “When it comes to AI, there’s a certain “black box” behind decisions that makes it so that even AI developers themselves don’t quite understand or anticipate the decisions an AI is making. We do know that neural networks are taught to make these choices by exposing them to a huge data set. From there, AIs train themselves into applying what they learn. It’s rather difficult to trust what one doesn’t understand.”
Drive.AI ready to add a layer of humanity to robot cars (Mashable.com)
Summary: “Carol Really wants to answer a question, “How do you get these robots that are driving on the road to interact with people?” The answer is simple. Along with all the 360-degree cameras, Lidar and a powerful Deep Learning AI engine, the car will include audio-visual cues to telegraph exactly what the car plans to do, what it sees and even give it ways to acknowledge pedestrians so they feel safer around these cars.”
Sense-Infer-Act-Learn: A model for trustworthy AI (O’Reilly Media)
Summary: “Rupert Steffner offers an overview of Sense-Infer-Act-Learn, a logical AI execution model to enable a more trustworthy AI. Sense-Infer-Act-Learn was inspired by various models, including the American jet fighter’s OODA loop, the cognitive science stimulus-response model, and the AI agent-reward approach to engineer a consolidated logical architecture model for real-time AI.”
Microsoft is Working to Make Artificial Intelligence More Human (Futurism.com)
Summary: “Microsoft’s Project Mélange, housed in its India office, is using code-mixing (moving between multiple languages within a distinct conversation or even a single sentence) to teach AI how to have more human-like conversations. India is a perfect location for this work because it’s a multilingual society in which many people are mixing languages regularly.”
What Are Some Of The AI Products You Can Start Building Right Now :
If you are a beginner it is recommended that you look up existing AI products to redesign and analyse them for ways to make them more trustworthy. Here are a few recommendations of how you can get started with building AI products:
1- Chatbots (Conversational AI)
Successful chatbots in different uses and industries: https://blog.appliedai.com/top-chatbot-success/
Chatbot Prototyping: 7 Tools to Shape A Bot With No Coding: https://chatbotsmagazine.com/chatbot-prototyping-7-tools-to-shape-a-bot-with-no-coding-f2bfac3c94e3
What is Conversation Design, and How to Design Your Chatbot
https://medium.com/swlh/what-is-conversation-design-and-how-to-design-your-chatbot-3754f04ab1e7
2- Artificial Intelligence Assistant like Siri
Build Your Own AI (Artificial Intelligence) Assistant 101(API.AI): https://www.instructables.com/id/Build-Your-Own-AI-Artificial-Intelligence-Assistan/
3- Smart Home systems/devices (Home Automation)
When IOT meets AI like Google Home, Alexa, etc.
https://www.instructables.com/id/When-IoT-Meets-AI-Home-Automation-With-Alexa-and-N/
4- Humanoid Talking AI bot
Something like Sophia?
Humanoid A.I Talking Robot With Arduino: https://www.instructables.com/id/Humaniod-AI-Talking-Robot-With-Arduino/
What Does It Mean To Be Designing For Trust? And “Reputation Capital” In Product Strategy.
The Currency of the New Economy is Trust
Summary: A TED Talk by Rachel Botsman: There’s been an explosion of collaborative consumption — web-powered sharing of cars, apartments, skills. Rachel Botsman explores the currency that makes systems like Airbnb and Taskrabbit work: trust, influence, and what she calls “reputation capital.”
Summary: A TED Talk by Joe Gebbia: Joe Gebbia, the co-founder of Airbnb, bet his whole company on the belief that people can trust each other enough to stay in one another’s homes. How did he overcome the stranger-danger bias? Through good design. Now, 123 million hosted nights (and counting) later, Gebbia sets out his dream for a culture of sharing in which design helps foster community and connection instead of isolation and separation.
Designing for Trust Observations from my first year at Airbnb (Airbnb.design)
Summary: “A year ago, I joined this company for the challenge of designing for trust. I don’t claim to be a social scientist or behavioral economist. I am a learner, and I believe in sharing ideas. So, here are some ideas I’ve thought of, heard, overheard, or otherwise acquired from a year of designing for trust at Airbnb alongside a team of the best designers, researchers, engineers, content strategists, and product managers I’ve ever known.”
Trust as Product Strategy (Forbes.com)
Summary: “Customers must be reassured that their data is in safe hands and that all efforts are being made to ensure it is used appropriately. Businesses need to remember that risks are not static, and are constantly evolving. They need to continually invest in their security, privacy, design, and be fast to deliver on customer expectations to build trust.”
The Trust Crunch (New Yorker)
Summary: “Systems of credit depends on trust. When trust is present, money flows smoothly from lenders to borrowers, allowing new enterprises to start, existing ones to expand, and daily business to move along without a hitch. When it’s absent, we find ourselves in a world where lenders hoard capital, borrowers are left empty-handed, and the economy’s gears grind to a halt — a world, in other words, like the one we’re now living in.”
(If you have recommendations on what else we can add to this repository, please leave a comment)
