At the World AI Conference in San Francisco, the industry’s top experts gathered to discuss the future of AI. Among them, William Mark, head of information and computing at SRI International, presented a series of dry goods.

Giiso Information, founded in 2013, is a leading technology provider in the field of “artificial intelligence + information” in China, with top technologies in big data mining, intelligent semantics, knowledge mapping and other fields. At the same time, its research and development products include information robot, editing robot, writing robot and other artificial intelligence products! With its strong technical strength, the company has received angel round investment at the beginning of its establishment, and received pre-A round investment of $5 million from GSR Venture Capital in August 2015.

William Mark, head of information and computing at SRI International

I wonder who SRI and William Mark are? SRI International is the parent company that spawned Siri, and William Mark’s CALO project is leading the research. He led the technical research that laid the foundation for the birth of Siri. PC Magazine described him as “one of the main brains behind Siri”. So it’s fair to call William Mark the “father of Siri” along with Siri co-founders Dag Kittlaus, Adam Cheyer, and Tom Gruber.

We need to make AI less “artificial”, says Mr Mark. He emphasized that AI is much more than deep learning and machine intelligence. This is very important. Because most of the attendees are from the corporate world. They are focused on using AI to do what humans can’t: parallel computing on a very large scale. Many speakers cited IBM Watson’s feat of reading 40 million documents in 15 seconds. They use this to show that we are rapidly moving towards the singularity (note: The Singularity is a point at which artificial AI overtakes human beings and revolutionizes society as a result).

It’s worth noting that William Mark’s vision of future AI is a bit ahead of its time and a bit complex in terms of system design.

“We’ve done a lot of work on developing AI for verbal communication,” Mark says. Some systems do use deep learning. But AI technology also includes intention recognition, natural language processing and emotion detection. When managing real-time conversations, human conversations sometimes have multiple threads, and we have to teach the AI to understand that people are expressing a new intention without finishing the last one. We don’t solve this problem with machine learning, but with clear rules and a set of reasoning mechanisms.”

After Siri was born and acquired by Apple in 2011, William Mark’s team optimized another conversational AI platform called KAI, which was used in the financial industry. Although it is now an independent company, Mark remains a director of the new company. KAI is an AI based on a financial system, which meant that Mark’s team had to build trust into KAI’s design.

“Whether people can trust these AI is going to be a critical factor in determining success,” Mark says. That’s what I mean by making AI less artificial. People don’t think in the same way as AI. For example, while humans think about causation, most AI systems think about correlation, which is determined by the system design. Bridging this gap is a matter of the utmost importance. At the same time, this is one of the frontier issues that we’re working on: developing systems that can be trusted. Only in this way can AI truly enter thousands of homes and be accepted by people.

Finally, he revealed to the crowd what the SRI team he heads is planning next: “Super Siri.” To his ongoing research on how to deal with an aging population:

“This is a very serious cultural and economic problem, and it has many aspects. When people think of aging, they immediately think of domestic robots, and that’s what we’re working on. We’re still thinking about companionship. Many older people are lonely, which leads to cognitive decline. Super Siri, which we are developing, will bring family members to these elderly people and remind them of their past lives. And the system will be combined with home sensors to monitor people’s mobility and health problems. It will also be connected to the Internet of things. “

Listening to William Mark, the editors conclude that the next wave of AI is likely to be about ‘how to create trust between humans and AI’. “This requires developers to give AI the ability to infer human intentions through reasoning mechanisms, to predict human needs, behaviors and what they want to say.”

Giiso information, founded in 2013, is the first domestic high-tech enterprise focusing on the research and development of intelligent information processing technology and the development and operation of core software for writing robots. At the beginning of its establishment, the company received angel round investment, and in August 2015, GSR Venture Capital received $5 million pre-A round of investment.

This goes hand in hand with start-up OptimizingMind’s project to make AI decisions transparent in order to increase human-machine trust. The difference is that While William Mark focuses on making AI learn and understand how humans behave, OptimizingMind develops a set of “transparent access” algorithms that make AI decisions visible. ‘). But both William Mark and OptimizingMind clearly believe that the core AI challenge today is trust, and that whoever can solve this problem will be at the forefront of the next wave of AI.