Cisco Spark Board: The Potential for the First AI-Powered Smartboard

One of the more interesting products showcased at the Nvidia Jetson TX2 launch last week was the Cisco Spark Board—a secure white board from Cisco which initially focuses on mostly making collaboration easy. Its initial core capability is that it blends messaging, video calling, and file sharing into a single product that is accessible from anyplace and secured and assured by Cisco. But I think this is only the start and when you couple an AI with a collaborative product like this, I expect it will evolve into a lot more.

Let’s talk about the potential future for the first smartboard, the Cisco Spark Board.

Facial Recognition

The NVIDIA Jetson TX2 was largely optimized to recognize objects. Core targets were autonomous vehicles that needed to find and navigate around things and security systems that needed—through facial recognition—to approve or alert based on who they saw. So, an obvious use would be to identify and authenticate the people attending a meeting. This would be both to assure that the intended attendees actually showed up and to alert if someone who was not invited was in the room to immediately address a security breach. This could have broader use as each person talking could be highlighted to the remote viewers for a meeting so they know who is talking. Currently it is nearly impossible to see name tags if you are remote and many people don’t wear them inside the company often making it hard for even local people to know who the person talking really is. This would be important not only for live events but for the record as a meeting taped and then played back years later could still showcase who the attendees were without someone having to try to match faces with the names of people who may no longer, at that time, work for the company.

Object Recognition

Despite the fact that both my mother and my step-mother were artists I can’t seem to draw to save my life. However, a white board product with the capability to recognize objects could guess at what I’m trying to draw on the white board and then, though a series of queries via text or voice, could determine where I was going and provide either a stock photo to work with or serve up a rendered image that would be far better than I’d have any hope of creating myself. This could work with everything from storyboards, to flow charts, to creating a presentation on the smartboard itself. It could play a major role in improving the quality of any live presentation collaboratively created. This would include words as well so that a chart with pictures and words would automatically morph into a something that could be indexed automatically based on content and appear far more polished than anyone but a true artist could create.

Visualizing Words

It is not uncommon to hire a graphics artists to attend meetings and create images that define what is being discussed to help with retention. A product like the Spark Board using the Jetson TX2 could listen to the talk and pull images from its memory or remote host computer to graphically showcase what was being said. This would be a deep learning implementation and require someone to create an app and related data to make this happen but this would not only help an audience better grasp what was being said but could also showcase to the speaker anything they were unclear or confusing on (because the system would be either alerting it doesn’t understand what is being said or drawing the wrong visual inference).

Wrapping Up: The Smartboard

Adding intelligence to a smartboard makes a ton of sense if only to make it drop dead easy to use. But I think that the Jetson TX2 has capability that isn’t being fully utilized yet that could transform this offering into substantially more. From using facial recognition to both create a better record for posterity and to identify unapproved attendees, to helping people become far better white board artists by taking over the creation, to converting words to images to help attendees to remember content—the use of intelligence in tools like this is barely off the ground but, I think, we could be on the forefront of something amazing!