July 19, 2018
Razorthink published in AI Informed

Interactivity and Artificial Intelligence

Artificial intelligence can bring interactivity into play between humans and machines – something early computer visionaries didn’t plan for. Mankind’s initial foray into the world of computational machinery was built on the idea of streamlining processes, but user error interfered, and constraints were designed to limit interaction.

A good example of a machine that is rigid in its accessibility to humans is the traditional ATM. While you have choices of which button to push and what data to enter, ultimately the desired results can only be achieved by following one form or another of the machine’s internal script – and deviation means failure.

Today’s machines are powerful enough to collaborate with humans and transform how we think about them. Artificial Intelligence and new types of interface make interactivity possible.  

Any interactive system takes the form of a cyclical process made of predictable steps:

  1.    Identifying a goal
  2.    Deciding to achieve the goal
  3.    Making a plan to achieve the goal
  4.    Executing the plan to achieve the goal
  5.    Identifying new information that may make it more difficult to achieve the goal
  6.    Accepting that the new information may change their plan to achieve the goal
  7.    Integrating the new information into their plan to achieve the goal
  8.   Adapting if necessary to identify a slightly revised goal
  9.   Repeat

If someone wants to eat ice cream, but on the way runs into traffic, making them late to pick up a child from school, or the ice cream place turns out to be closed, or it becomes extremely cold and ice cream is no longer as attractive an option, the cycle interrupts and may rest to an earlier step.

When the goal depends on other humans, adaptation is easy – modeling behavior comes into play, and behaviors change based on interaction and understanding, leading to cooperation, collaboration, negotiation, and relationships.

Most machines don’t model users and don’t adjust their behavior, instead expecting the user to handle all the goal related steps themselves and only communicate with the machine via its own strict parameters. The machine remains in control – it doesn’t care if the user succeeds or not.

Interactive AI systems can use data on the human to create an interactive experience, with predictive behavioral analysis and adaptive responses. Machines powered by artificial intelligence will become capable of modeling the user, observing the user’s actions, and extracting meaning. A highly developed AI will be able to understand not only who the user is, but what they want, when they want it, and dynamically adjust to new user behavior and other inputs  and assist the user in achieving their goal instead of merely facilitating through a rigid system.