Have you ever wanted to tell a robot what to do using your own words, like you would to a human? Well, thanks to Microsoft’s latest research, now you can – with the help of its ChatGPT AI program.
Monday, Microsoft’s researchers published a paper demonstrating how ChatGPT can streamline the process of programming software commands to control various robots, such as mechanical arms and drones – making it much easier to have your robot do exactly what you want it to do.
Microsoft’s approach taps ChatGPT to write some of the computer code. How does ChatGPT do this? It was trained on huge libraries of human text, including the code for software programs. ChatGPT has already shown it can write and debug programs in various languages based on text-based requests, making it ideal for robotics hardware.
To help ChatGPT write the computer code, the researchers first outlined to the AI program the various commands it could use to control a given robot. The AI chatbot was then given a prompt to begin writing the computer commands for that robot.
One example of a prompt given to ChatGPT was to control an aerial drone. In this example, Microsoft researchers fed the AI chatbot a rather long prompt laying out the computer commands it could write to control the drone. After that, the researchers could make requests to instruct ChatGPT to control the robot in various ways. This included asking ChatGPT to use the drone’s camera to identify a drink, such as coconut water and a can of Coca-Cola. Remarkably, ChatGPT was able to interpret the request and write computer code for the drone to fly in front of a mirror and take the selfie.
In another demo, the researchers used ChatGPT to write code capable of directing a robot arm to build the Microsoft logo using several wooden blocks. The ChatGPT AI model was able to write the code with ease, after being given the initial prompt.
Of course, the ChatGPT approach still has a key limitation: the chatbot can only write the computer code for the robot, based on the initial “prompt” or text-based request the human gives it. Hence, a human engineer has to thoroughly explain to ChatGPT how the application programming interface for a robot works, otherwise the AI program will struggle to generate applicable computer code.
This breakthrough from Microsoft does open the possibility of using ChatGPT for more complex tasks as well. For instance, a human could request for ChatGPT to control a robot to complete a specific task based solely on reading a series of instructions, such as having a robot move a mirror in order to take a selfie. In this case, the AI chatbot would have to interpret the instructions, figure out what the task entails, and then write the appropriate computer code.
ChatGPT is an incredible achievement that has the potential to revolutionize robotics, as well as many other areas of AI. So why not see what this amazing AI program is capable of? After all, with ChatGPT, you never know what it will come up with next.
Whether you are a roboticist, an AI enthusiast, or just someone who is curious and wants to explore what ChatGPT is capable of, Microsoft has released an open-source platform on GitHub, where anyone can share examples of prompting strategies for different robotics categories.
When using ChatGPT to control a robot, it’s always best to be mindful of the potential risks associated with AI programming. An incorrect response from ChatGPT can lead to safety-critical issues when it comes to robotics, so be sure to ensure solution quality and safety with human supervision before executing code on the robot.
By tapping into the power of ChatGPT, Microsoft is making it easier and faster to have robots do the tasks you want them to do. Make sure to stay tuned and see what else Microsoft has in store with ChatGPT. Who knows, you might just be in for some amazing surprises!