If you are like me, you might have noticed that AI has become a part of everyday life. I wake up every morning and ask my smart assistant about the weather. Recently, I applied for a new credit card and the credit limit was probably determined by a machine learning model. And when I typed the previous sentence, I got a word choice suggestion that “probably” might sound better than “maybe”, an AI-based suggestion.
As a member of Google’s Responsible Innovation team, I think a lot about how artificial intelligence works and how to develop it responsibly. I recently spoke with Patrick Gage Kelly, Research Strategy Lead for Google’s Trust and Security team, to learn more about developing products that help people recognize and understand artificial intelligence in everyday life.
How to help people navigate the world with so much AI?
My goal is to make sure that people at a basic level know how artificial intelligence works and how it affects their lives. AI systems can be really complex, but the goal of explaining AI is not to make everyone become a programmer and understand all the technical details, but to make sure that people understand the parts that are important to them.
When artificial intelligence makes a decision that affects people (whether it’s recommending a video or eligibility for a loan), we want to explain how that decision was made. And we don’t just want to provide a complicated technical explanation, but rather information that is meaningful, useful, and helps people take action when needed.
We also want to find the best time to explain AI. Our goal is to help people develop AI literacy early on, particularly in primary and secondary education. And when people use products that rely on AI (everything from online services to medical devices), we want to give people many chances to learn about the role of AI, as well as its benefits and limitations. For example, if people are told at an early stage what mistakes AI-based products can make, they will be better prepared to understand and correct situations that may arise.
Do I need to be a mathematician or programmer to have a full understanding of AI?
No! A good metaphor here is financial literacy. While we may not need to know every detail of what’s involved in an interest rate hike or the intricacies of financial markets, it’s important to know how they affect us – from paying off credit cards to buying a home or paying off student loans. Similarly, the explanability of artificial intelligence is not about understanding every technical aspect of a machine learning algorithm, but about how to interact with it and how it affects our daily lives.
How should AI practitioners – developers, designers, researchers, students, and others – think about AI explainability?
Many practitioners do important work on explainability. Some focus on interpretability, making it easier to identify specific factors that influence decisions. Others focus on providing “instantaneous explanations” just as the AI is making a decision. These can be useful, especially if carefully planned. However, AI systems are often so complex that we cannot rely entirely on snapshot explanations. There is simply too much information to pack into an instant. On the contrary, AI education and literacy should be incorporated into the entire user journey and developed continuously throughout a person’s life.
More generally, AI professionals should think about the fact that AI explainability is fundamental to the design and development of the entire product. At Google, we use AI principles to guide responsible technology development. In line with Principle #4: “Be accountable to people”, we encourage AI professionals to think about all the points and ways they can help people understand how AI works and makes decisions.
How do you and your staff work to improve AI explanations?
We develop resources that help AI professionals learn creative ways to build AI explainability into product design. For example, in the PAIR Handbook, we launched a series of ethical case studies to help AI professionals think through difficult questions and hone their AI explainability skills. We also conduct basic research like this article to learn more about how people perceive AI as a decision maker and what values they would like AI-based products to embody.
We’ve learned that many AI professionals need concrete examples of good AI explanations that they can build on, so we’re now developing a story-centric visual design toolkit for explaining a fictional AI application. The toolkit will be publicly available, so teams at startups and tech companies around the world will be able to reference the explanation in their work.
I want to learn more about the possibility of explaining AI. Where to start?
This February we released the Applied Digital Skills lesson “Discover AI in Everyday Life“. This is a great place for anyone who wants to learn more about how we interact with AI every day.
Do you want to consult or try something from Google services? Send your requests to info@wiseit.com.ua or leave a message in the chat on wiseit.com.ua – we will contact you!