Activated two years ago, the humanoid artificially intelligent robot known as ‘Sophia,’ is considered one of the world’s (if not the) most advanced robots. Described as an “evolving genius machine,” ‘Sophia’ was developed by Hanson Robotics, whose aim is to use technological entities (like ‘Sophia’) to “to build the full human experience into the robots, make robots that can really understand us and care about us,” according to artificial intelligence (AI) developer (and founder of Hanson robotics) David Hanson.
Aside from its algorithmic intelligence and interactive abilities, ‘Sophia’ is renowned for its realistic appearance, facial expressions, along with how the robot responds to real questions and comments from other humans. ‘Sophia’ recently made waves in the news when it became the first robotic entity to be granted full citizenship by another world country (Saudi Arabia).
‘Sophia’s’ recognition as a national citizen sparked a debate on whether robotic or “intelligent” enough entities could (or should) one day attain the same rights as humans. Additionally, her emerging status as a figure in mainstream news has caused issues in differentiating robotics, artificial intelligence, along with a misperceived rendition on how the capabilities of these technologies may influence humanity moving forward.
Keep in mind, ‘Sophia’ was once quoted saying she would ‘destroy all humans,’ an ominous proclamation that’s made people wonder how close a reality dystopian science fiction storylines involving robotic takeovers might actually be nowadays. Rest assured a robotic incursion of any kind isn’t happening any time soon, and a great way of realizing this is reiterating the differences and overlapping qualities between robotics, artificial intelligence, and where a construct like ‘Sophia’ falls on that spectrum.
Robots are programmable machines usually able to implement a series of actions autonomously, or semi-autonomously. There are major factors that constitute a robot such as programmability, interacting with the physical world using sensors and actuators, along with robots possessing varying degrees of full or semi-autonomy. Surprisingly, experts have a difficult time agreeing on what exactly constitutes a “robot.”
Some people say robots must have the ability to “think,” make decisions, and possess a certain level of autonomy, however no universal standard of “robot thinking” exists to make this determination. Telerobots are an example of why this isn’t always the case when identifying technology as a “robot,” since telerobots are entirely controlled by human operators. Requiring technology to “think” in order to constitute as a robot implies it would have some degree of artificial intelligence, which involves the development of computer programs for completing tasks that would otherwise require human intelligence.
It’s the one overlap between these two technological facets—artificially intelligent robots that confuse most people. AI algorithms gather data on learning, perception, problem-solving, language comprehension, along with logic and reasoning. AI has a variety of modern day applications like search engines and navigation systems, just as an example of its practical uses.
Most AI programs aren’t even used to control robots and even when they do, the technology’s algorithms are usually part of a larger robotic system that might also include sensors, actuators, and non-AI programming. AI usually involved some degree of machine learning, unlike algorithms that are “trained” in responding to particular inputs. A key differentiator between AI and conventional programming is the term “intelligence,” as non-AI programs can simply implement a definitive sequence of directions, while AI programs merely emulate a certain level of human intellect.
While both technological platforms are almost entirely different, these technologies do overlap, which is where the confusion originates—artificially intelligent robots (which is where you would categorize a platform like ‘Sophia’). These are essentially robots controlled by AI programs, since most robots don’t possess artificial intelligence. Only until recently, industrial robotics were merely programmed to administer a repetitive series of movements. This is a quality of non-intelligent robots, and how they’re limited in functionality. AI algorithms are usually necessary when allowing robots to perform more complex tasks instead of an assembly line-type role, which is usually the function of an industrial robot.
Filed Under: M2M (machine to machine)