A technical person is required to handle a robot because robots communicate with codes, which necessitate programming skills. ChatGPT can be used here to control them using our own words, like we would a human, rather than hand-written control codes.
ChatGPT has astounded everyone with the capabilities it brings with it. Ever since its release, new avenues and use cases have opened with the innovative application of the language model. In one such case, Microsoft is experimenting with ChatGPT controlling robots to make them understand human-like commands.
In a viral video posted on the microblogging site Twitter, a person is seen posing questions to a robot that, when integrated with ChatGPT (claimed in video), comprehends prompts and gives human-like answers. Thus, ChatGPT gives robot an anthropomorphic communication qualities.
Why robots need technical person to train it?
A technical person is required to handle a robot because robots communicate with codes, which necessitate programming skills. The usability of robots can drastically increase if we can control them using our own words, like we would a human, rather than hand-written control codes. Microsoft’s scientists believe that using OpenAI’s new AI language model, ChatGPT, they can achieve natural human-robot interactions.
“Our goal with this research is to see if ChatGPT can think beyond text, and reason about the physical world to help with robotics tasks. We want to help people interact with robots more easily, without needing to learn complex programming languages or details about robotic systems,” a blog post from Microsoft read.
How can ChatGPT transform the robotics industry?
In the present scenario, robot training needs an engineer to translate the task into code for the machine. The engineer needs to remain in the loop when correcting the robot’s behaviour by writing new codes. Overall, this process is slow, expensive, and inefficient, Microsoft says.
But now with ChatGPT in play, it allows a non-technical user to sit on the loop and instead provide feedback to the large language model (LLM), which in turn writes the code for robots, the report adds.
Members of the Microsoft Autonomous Systems and Robotics Research Group tested ChatGPT under various conditions. The model was asked to create an algorithm for a drone to reach a goal in space without colliding with objects, and it was tested to see if it can decide robot route in real time based on sensor feedback. ChatGPT has also been used in a robot arm manipulation scenario, industrial inspection, and testing the ability to operate a real drone by the tech giant. These experiments yield encouraging outcomes.
“We believe that language-based robotics control will be fundamental to bring robotics out of science labs, and into the hands of everyday users,” the report says.
ChatGPT, which is a language model trained to understand human-like prompts and questions, has to receive further training to make these possible. The challenges that need to be tackled are teaching ChatGPT how to solve problems considering the laws of physics, the context of the operating environment, and how the robot’s physical actions can change the state of the world.