Cognitive Assistance: Not Your Killer Robot AI

Artificial intelligence is a hot news topic lately. Headlines read like a spine-chilling science fiction movie: “Artificial Intelligence is as dangerous as NUCLEAR WEAPONS': AI pioneer warns smart computers could doom mankind” or “Scientists Warn that Robots and Artificial Intelligence Could Eliminate Work.” Yikes! This view of AI can seem dystopian and dehumanizing, yet the reality of machine intelligence today is something quite different.

Today, AI systems are commonly used for sifting through lots of data to make trained predictions. Just about every tech titan uses deep learning or other predictive tools to grow their business. The key word here is predict! Deep learning has gotten very good at very specific tasks with massive datasets to help us predict if our friends are in an image or if we will like a certain piece of content online.

These early AI use cases are just the beginning. Our personal and work lives are becoming more and more data rich, and making sense of it all is getting exponentially harder. We have an opportunity for AI systems to perform some of the heavy lifting. Rather than AI being viewed as a human replacement, maybe it’s time we saw it for what it is — an assistant.

Halfway through writing this post, a brilliant Watson ad spot popped up on the screen. The commercial features Carrie Fisher (best known for her role as Princess Leia) leading a support group for robots that need help “coping with humans.” Unlike the other sci-fi robots in the group, Watson is represented as a talking monitor that enjoys working with humans to help them solve problems together. Carrie Fisher emphasizes the “with humans” line. Sure, this is a marketing ploy to sell AI software but it embodies our mission — AI assisting us in our daily lives.

Cognitive assistance is about combining the computing resources of machines and the contextual understanding of humans. Our synaptic intelligence platform provides not an answer, but recommendations for humans to evaluate. To help them, we provide explicit "why" reasons with each recommendation. We’re replacing rules-based and statistical models that rely on constant programming or brute force data processing to a new form AI focused on how data connects to other data, inspired by neuroscience and how our brain connects information.

For an introduction to some of the problems we are solving with synaptic intelligence, see our presentation at Connected Enterprise.

Constellaton's Connected Enterprise - Nara Logics