AI and machine learning promise a plethora of benefits to society in general – the trick is convincing society that they can trust AI-powered gadgets and services. Doing that requires plenty of public education on what AI can and can’t do, as well as ensuring that decisions requiring moral or value judgments are left to humans.
Those were the main themes of a panel session in Singapore last week – hosted by mobile classifieds marketplace Carousell, and featuring speakers from Google and Microsoft – on the transformational impact of AI and machine learning.
Unsurprisingly, everyone on the panel agreed that AI and machine learning will play a major role in developing more efficient technologies and services in the near future, although that role will be much clearer further down the line than it is now.
“I firmly believe that everyone in technology, say five years from now, will have to have a good understanding of machine learning and AI, as it will be a part of almost every application that is out there,” said Oyvind Roti, head of Solutions Architecture (APAC and Japan) at Google Cloud. “Inside Google, it already is.”
One potential barrier to that understanding is that many people still have misconceptions about AI – in part because they associate AI with blockbuster movies and TV shows, or because of comments from high-profile tech gurus like Elon Musk. Consequently, both businesses and media need to do better in educating the public about the benefits of AI – and tech companies in particular need to find a way to communicate that to people who aren’t that tech-savvy, said Oyvind.
“We are finding ways to explain some of the more complex concepts around AI without getting all caught up in the math. It’s important to us to lower these barriers and accelerate how researchers, developers and companies work in this field,” he said.
Chris Auld, principal technical evangelist manager for Southeast Asia at Microsoft, added that while AI can help to make communities safer, human interaction, collaboration and policing is still essential.
“Human judgment remains really important when it comes down to making value judgment as it’s very hard to imbue machines with values or morals,” he said. “So AI and machine learning can help us narrow down the set that humans need to look at so that they can put more effort and more thought into that value judgment.”
Carousell co-founder and chief technology officer Lucas Ngoo said that his company does this to police its marketplace community and protect users from scammers, fraudsters and other bad actors.
“We rely on AI and machine learning to help us predict behaviors of bad actors on the platform as a first line of defense,” he said, “but while we’re deploying AI tools, we also depend on user reports and flags to help make Carousell safe and secure for everyone.”
Ngoo said that in Singapore, fraud cases have fallen quarter on quarter and make up less than 0.05% of all transactions on Carousell. “By and large, we have actually seen the fraud situation improve on our marketplace.”
The overall goal of this is, of course, to build trust with customers – not just trust in the platform, but in the AI tech that governs it. Microsoft’s Auld commented that trust is a crucial factor as the digital world becomes more integrated into our daily lives. People in online communities need to feel those communities are safe and trustworthy – and ensuring that is an industry-wide challenge that requires collaboration across companies, industries and governing bodies, especially in the technology space.
“The things we can do with AI and machine-learning far outstrip the things that the community is comfortable with us knowing,” Auld said. “The responsibility is on us to approach it slowly and carefully to build trust over time, so we can take the community on the journey to use the communities’ data responsibly and improve things on the whole.”