You've probably heard a lot about AI and robots. People often talk like they are the same thing.
But, there are a lot of myths out there about robots and even more about AI.
Here are 5 common myths and the truth behind them.
If you've read a lot about Artificial Intelligence (AI) and robots you might be a bit overwhelmed by the amount of differing information available out there.
Some sources talk about AI as if it is already being applied to all levels of business — and that our jobs are all at risk as a result — while others proclaim that AI is "the future" and we all need to invest in it for our business right now.
Then, journalists sometimes mix up the concepts of robotics and AI when they write about them. An article or video will proclaim "AI may replace engineers" but the image nearby is of an industrial robot which almost certainly has never been programmed by AI.
As we've said before, there is a clear difference between the concepts of AI and robotics. But, even if you know that, it's not always clear what information you should believe.
The problem is that there are so many articles written about AI and robots every day across the internet. More words are written than the amount of real, new information available.
Technology journalists and bloggers are keen to show that they are writing "on the cutting edge" by writing about the latest robotic developments week after week.
But, the reality is that the robotics industry changes very little each year. Most years, there are only ever a small handful of noteworthy new developments in the industry. The rest are minor tweaks to existing AI and robotic technologies.
Here at Robotiq, we handle this problem by spending most of our time writing about useful content that will actually be helpful to you. However, many outlets fill the "lack of new information gap" by making copious predictions about "the future." These are usually sensationalist and tend to include a lot of supposition about how robots and AI will advance in the next few years.
This trend is where many of the most pervasive myths have sprung from.
Here are 5 common myths about AI and robots, followed by the truth behind them.
There is an assumption among many that robots are programmed by artificial intelligence. People think that robots are capable of making decisions and planning their own actions.
The truth: In reality, very few robots are programmed by AI. This belief stems from the general confusion between the two terms. As explained in our article on the topic, the crossover between robots and AI (i.e. artificially intelligent robots) only constitute a very small amount of robots.
In industrial robotics, the use of AI programming could potentially be quite dangerous. A robot with the capability to plan its own movements on-the-fly is naturally unpredictable. In most cases, it much more preferable to program all the movements before running the robot.
The myth that AI is already superintelligent can often be seen in the media. Self-learning robots skilled at particular tasks are a large driver to the myth, such as Google's AlphaGo which beat the world champion at the game Go.
The truth: The truth with all AI (not just AI robots) is that any AI is only ever good at one particular type of task. To my knowledge, there is no reliable general-purpose AI suited to all tasks and there is unlikely to be for many years (if ever). For example, AlphaGo has an exceptional ability to play Go but it is useless at every other task.
Impressive AIs are always highly specialized and, as a result, are not intelligent in the ways that we humans usually think intelligence.
Some people believe that an AI robot will be able to handle itself and solve any problems on its own. They believe that having AI means that humans workers do not have to do anything and the robot will do everything.
The truth: No AI can function without human input. At the very least, an AI robot will always require you to set up the workspace to aid its functioning.
A good example of this is the Roomba vacuum cleaner now found in many homes. It can autonomously navigate a home using a form of AI to map the floorplan and plan its route. However, if you leave cables and junk on the floor, it can easily get stuck. The human's job is to give the AI robot optimal conditions so it can do its job.
Just as people think that AI robots can program themselves (they can't) people often think that you can just leave an AI robot alone and it will function indefinitely. The robot has intelligence therefore it is intelligent.
The truth: As you've already seen in some of the previous myths, this is just not true. AI is not intelligent in the way that humans are intelligent. They have high performance for a specific type of task. They will probably always need people — even if it's just to ensure that they have not deviated from their task.
For those of us who know a little bit about artificial intelligence, there is often a belief that AI requires extremely large datasets to function. For example, an AI robot might require thousands of images of the same product to be able to detect it correctly.
The truth: When it comes to data, a good rule of thumb is that you only need as much data as you need to get a good result. Our Wrist Camera uses template matching, a very basic form of AI, but can reliably detect a product with a single training image as a dataset. Similarly, not all AI algorithms require many data points.
Whichever myths you hear, just remember that AI is a specialized tool that can improve robotics in some specific situations but is still not widespread. Don't believe the myths.
Which myths have you heard about AI and robotics? Tell us in the comments below or join the discussion on LinkedIn, Twitter, Facebook or the DoF professional robotics community.