02: Chatbots and Social Identities Part 2

“Many times we try to think of how we can replicate the formal process of learning into the AI, which can be a challenge.” - Mo

(click to tweet)

We return with part two, for the continuation of our discussion on the current and future world of AI. We are discussing the advances toward a general intelligence in AI, and how scientists are working to develop a whole new type of unsupervised learning for the surrogate to take on.

• Momin Mirza: Verizon corporate strategy group based in Palo Alto, focused on trending topics, particularly AI and commerce

• Esther Crawford: CEO and Co-Founder of Olabot. Their goal is to make it fun and simple for everyone to have a personal bot

• Raj Ramaswamy: CEO and Cofounder of ShopInSync, whose goal is to enable brands and retailers in the commerce space to build deep contextual bots that help get their conversions up and drive more revenue for them

• Mounir Shita: Founder and CEO of Kimera, with the first AI that comprehends what it’s learning and to automate any type of connected device and network

“You want to be able to trust your devices to do the right thing.”

(click to tweet)

Highlights

• Computers have to go through an enormous amount of training in order to gain intelligence
• Scientists are currently focused on making true “unsupervised learning” a reality
• In the field of general intelligence, there is no such thing as “training data”
• Deep learning by itself is not going to lead to general intelligence
• It will most likely require a whole new way of accomplishing AI to create true general intelligence
• The advantage of machines over humans is the different set of sensors machines have
• Machines can learn about their environment from a different perspective than we can
• We will have to trust the AI to do tasks on behalf of us. This is a significant leap for most people
• Comfort and trust is growing over time in the general population to let machines and devices make decisions for them
• We’re never really 100% honest when we talk about ourselves to the world, but we potentially could be with a bot
• Some bots are trained with continuous conversation
• Bots can be thought of as a new model for communicating with users

Resources

http://www.shopinsync.com/

http://olabot.com/

www.kimera.ai