AI in Industrial Robotics
, by Jim Ryan, 1 min reading time
An examination of the possible future role of AI in manufacturing by Zach Boyd at the Hirebotics blog. For instance, "When applied to welding, AI systems can fine-tune their own parameters for speed, heat, and angle of welding, making micro adjustments where necessary."
To review: The terms "automation," "robot," and "artificial intelligence" aren't always clear to everyone at all times. One way to look at it is that they demarcate increasing levels of autonomy in machines. Simple automation can be achieved without programming the machine to make any decisions. Robotics introduces decision making based on sensation. Robots have sensors and programs that make decisions based upon the data received from them. The decisions can become somewhat undetermined by just the program alone as the program takes more and more environmental factors into account. (Or, at least, the robot's decisions can become practically unpredictable by humans examining the code.) These are all increases in autonomy. AI introduces the ability of the robot to enhance its program by learning from experience: a further increase in autonomy.
In light of all that, Boyd's point about welding - "fine-tune their own parameters for speed, heat, and angle of welding, making micro adjustments where necessary" - can be classified as mere robotics and not AI, depending on the manner in which the robot makes these adjustments. A mere robot can be programmed to make such adjustments, but if they require the robot to learn new things, then the robot is at the level of AI.