AI should be Augmented intelligence.

I know what you are thinking. Technically, Augmented Intelligence is different from Artificial Intelligence (AI). It’s just a sub-form of AI. Check! But that’s not my point.

Most practical and realistic (and successful) AI initiatives has been focused on augmenting human abilities or intelligence with powerful machines. And I believe it’s because most folks intuitively cannot think about what machines can do that humans haven’t already attempted.

Given we are coming off the #AI winter, there are a lot of folks who I’d call as the AI police who have been and they are beginning to feel a little loss of identity, but truth is that we all have to get over it, as I’m convinced that it’s going to be summer for some time, especially if we focus our efforts on “augmenting” human intelligence than trying to replace it!

It’s simple, human cognitive capabilities have limits, and if you can use machines to augment that, you make these humans more valuable!

Let’s take an example to make that point:

Doing real time or even delayed security footage/video analysis to identify objects, people and or actions has a huge role to play if the machines can do the work for you. It won’t replace the security guard or make her lose her job, but instead she will now be able to focus on taking the right action if the machine identifies a threat, like an intruder. I am not sure we are comfortable (yet) to have the machines act on the behalf of human in such situations (unless in a war setting may be). But in this example, you just augmented the security guard’s abilities to identify intruders with a high degree of accuracy and take necessary action. She just became a lot more effective.

This kind of pattern/facial/object/whatever recognition through video vision is a very standard and leverages several popularly available (deep learning) packages or cloud APIs (AWS/Google/Azure etc.) and can be implemented easily, given enough computing horsepower. Just a few days ago, I saw this Microsoft announcement of Azure Cognitive Services.

Azure Cognitive Services — Will enable users to infuse apps with cognitive intelligence. Face API, Computer Vision API and Content Moderator are available in the Azure Portal. With these updates, Microsoft will have 25 APIs and services in their Cognitive Services collection.

Source: Kipod.ca — screenshot, human tracking

Kipod, a Canadian startup seems to have a super cool solution for this exact use case: https://kipod.ca > check them out. Here’s a screenshot from them.

For a more business context, Enterprises should strive to be a self-driving cognitive business, which not collects/analyzes and stores data from their customers, partners, employees and markets, but has a cognitive way to interpret, recommend and act on it. After all, it’s all about being more human in the way businesses act.

Lastly, I don’t think as humans we may want to emulate the intelligence of any other being, living or non-living, but there may be a (hint: inter-galactic) point of leaving that option for a future where humans are not just the only truly intelligent species.

PS: After I wrote this, I did a quick search (duh!) and found that there is a growing consensus on this topic.

http://www.informationweek.com/government/leadership/ibm-ai-should-stand-for-augmented-intelligence/d/d-id/1326496

https://venturebeat.com/2017/01/09/why-ai-must-be-redefined-as-augmented-intelligence/