Chatbots are computer programs that process and simulate a human conversation… and they are in operation everywhere. In 2022, 88% of web users interacted with a chatbot.
They help us order our burritos, suggest products we might like, debug code, synthesize complex information, create visual art, and even offer medical advice. Sonny Patel, VP of Product Engineering and Conversational Intelligence at LivePerson, recently joined Round to discuss how chatbots — from Sephora’s virtual beauty coach, Ora, to ChatGPT — are changing how people interact with the world around them.
According to Tidio, 62% of people would rather interact with a bot than a human in common customer-service situations, like ordering takeout, checking a delivery status, or shopping online. Chatbots excel at these routine tasks: they work 24/7 to interpret human language, search vast data stores, and generate rapid, useful responses. But in areas like trouble-shooting or product returns, polling shows that chatbots still leave something to be desired — this may be changing though. In 2022, ChatGPT became the second AI to pass the famous Turing Test, signaling rapid innovations in the field. As technology advances, Patel predicts sophisticated chatbots will soon draw from account profiles and past conversations to create an even more human-to-human-level experience for consumers.
Chatbots benefit businesses in a few clear ways. They allow brands to rapidly reply to customers around the clock, increase lead generation, and cut costs by handling repetitive questions. But chatbots also create a valuable feedback system with the customer. Brands no longer have to rely on indirect measures like abandonment and click-through rates to understand consumers’ experiences on their apps and websites. “If the customer isn’t happy with something,” Patel says, “we can have a conversation and ask: How can I make your experience better?”
Much like the advent of Google search, AI advancements will transform the way we work. As chatbots grow more capable and versatile, they will become indispensable in a range of tasks and industries, from finance to healthcare. But rather than eliminate jobs, Patel predicts chatbots will make us better at the work we already do. One day soon, we may all have personalized, industry-specific AI coaches who enhance our efficiency and performance.
Given rapid innovation in AI, industry players must work now to address bias in machine learning — and this means pushing for governance around AI’s role in day-to-day life. Today, AI’s decision-making processes are often opaque and difficult to understand, making it hard to address concerns about bias and discrimination. But what if all chatbots could explain the factors behind their outputs in human terms? That would give individuals the power to understand why an AI gave certain suggestions or responses. In 2019, for example, Apple co-founder Steve Wozniak tweeted about an instance of algorithmic bias where he was given 10x more borrowing permission on his Apple Card than his wife even though they shared all their assets and credit cards. “Building explainable AI is powerful in these situations, Patel says, “because it allows us to access an AI’s decision-making criteria and say: Why did you consider this or that aspect in your decision? That doesn’t seem right.” In our polarized world, tackling machine bias is a challenge, but Patel says we can start by embracing the push for explainable AI.
Every designer working today is trying to make the world a better place, yet it is often forgotten that we are not only designing for people but also for the planet.
Dan Van Tran (DVT) is the CTO of Collectors Holdings, a leading player in the collectibles industry. Tran talked through one of the most difficult challenges tech and product leaders must solve: building transformational technology on top of aging technology.
Building responsible and equitable algorithms is one of the toughest challenges technologists face today. Expert Cathy O'Neil explains why and what to do about it.