Intelligent robot turning point? Microsoft will use ChatGPT to control robots

The powerful functions of ChatGPT have made many technology companies want to explore its application boundaries. For example, what kind of surprises can it bring us by giving it a robot body?

A few days ago, Microsoft published a technical article on its official website, introducing that Microsoft is trying to let ChatGPT break through computer limitations to control robots so that they can play a role freely in the real world.

According to the article, Microsoft will expand ChatGPT's capabilities to the field of robotics and use language to intuitively control multiple platforms, such as robotic arms , drones , and home assistant robots.

For example, if you say to your home robot, "Please heat up my lunch," the robot will then find the microwave and heat up the food. This is obviously our ideal robot. We can tell the robot what to do with our own words, just as easy as communicating with other people.

Language is the simplest and most intuitive way for us to express our intentions, but now we can only rely on writing code to control robots, which undoubtedly makes the operation much more complicated. Currently, the Microsoft team is exploring how to change this reality and trying to use OpenAI’s new AI language model ChatGPT to make natural human-computer interaction possible.

Application of ChatGPT in robots

ChatGPT is a language model trained on a large corpus of text and human interactions. It is able to generate coherent and grammatically correct responses to a wide range of prompts and questions. To use ChatGPT to control a robot to complete tasks, it is necessary to enable it to think beyond text and deduce the physical world.

"The goal of the research is to see if ChatGPT can go beyond generating text and reason about real-world situations to help robots complete tasks," Microsoft researchers said. Microsoft hopes to help people interact with robots more easily without having to learn complex programming languages ​​or detailed information about robotic systems.

Current robotics use is based on engineers or technical users who can translate task requirements into system code. Engineers need to frequently write new code and specifications to correct the robot's behavior. This process is slow (users need to write low-level code), expensive (requires highly skilled users with in-depth knowledge of robotics), and inefficient (requires multiple interactions to make things work properly).

ChatGPT unlocks a new robotics paradigm that can quickly convert natural language into code, so that a non-technical person can provide high-level feedback to the Large Language Model (LLM) while monitoring the robot's performance. Based on this, Microsoft uses the knowledge of LLM to control robots of different shapes to complete various tasks without any fine-tuning.

It is understood that the current experiment has been able to input instructions into ChatGPT's dialog box to control the robot to find "healthy drinks", "things with sugar and red signs" (in this case, a can of Coke), and a mirror for the drone to take selfies in the room.

It turns out that ChatGPT itself can indeed do a lot, but some technical improvements are still needed, including but not limited to special prompt structures, advanced APIs, and human feedback through text. Microsoft believes that their current efforts are just the beginning of a shift in the way robot systems are developed, and hopes that more researchers will enter this exciting field.

ChatGPT brings inflection point to robot applications

Using robots to help us deal with various trivial matters in life may sound like a science fiction idea, but in fact it is not so far away from us, or it has begun to enter our lives. For example, the sweeping robot has become the most widely sold and used robot in the world.

A few days ago, a study published in the journal PLOS ONE caused heated discussion. The study invited 29 artificial intelligence (AI) experts from the UK and 36 AI experts from Japan to predict and analyze the role of AI in the future family field.

Experts generally believe that in 10 years, about 39% of housework will be completed by automated programs. Of course, the degree of robot intervention will vary for different housework. For example, the task of grocery shopping (mainly daily necessities) may be most affected, and the time people spend on this will be reduced by about 60%; while the task of taking care of children may be the least affected, and the time people spend on this will only be reduced by about 20%.

The core technology modules of robots can be simply divided into three aspects, namely perception system, intelligent AI system, and motion control system. If ChatGPT can be successfully applied to robots, the intelligent AI system will achieve a breakthrough, which means that we can communicate smoothly with robots and use language to control robots to complete various tasks.

In fact, some organizations have pointed out that ChatGPT may bring the next turning point for humanoid robots. As a multi-tasking, general-purpose robot platform, artificial intelligence similar to ChatGPT is expected to further open up its application scenarios in industrial production, social services and more fields.