AI and humans collaborate on first ChatGPT-designed robot

Is there anything that ChatGPT can’t do? Yes, of course, but the list is getting shorter and shorter. Now, researchers have used large language models to help them design and build tomato-picking robots.

Large language models (LLMs) can process and internalize large amounts of text data, using this information to answer questions. OpenAI’s ChatGPT is one such LLM.

In a new case study, researchers at Delft University of Technology in the Netherlands and the Swiss Federal Institute of Technology (EPFL) enlisted the help of ChatGPT-3 to design and build a robot that, strange as it may sound, That ChatGPT is a language. Sample.

“Even though ChatGPT is a language model and its code generation is text-based, it has provided important insights and intuition for physical design, and has shown great potential as a robust board to stimulate human creativity,” said Josie Hughes, co-author. Published case studies about the experience.

First, the researchers asked the AI ​​model, “What are the future challenges for humanity?” ChatGPT proposed three: food supply, growing population and climate change. The researchers chose food supply as the most promising direction for robot design because it was outside their area of ​​expertise.

Using the LLM’s access to global data sourced from academic publications, technical manuals, books and the media, the researchers asked the AI ​​what characteristics a robotic harvester should have. ChatGPT came up with a motorized gripper to pull ripe tomatoes from the vine.

Once this general design is decided upon, researchers can move on to the specifics of the design, including the construction materials that will be used and the computer code that controls it. Currently, LLMs cannot create complete computer-aided design (CAD) models, evaluate code or automatically build a robot, so researchers need a ‘technician’ for this step. required to take on a role where he assisted with these aspects, optimizing code written by LLM, CAD Finalization and Robot Fabrication.

“Computation has been used extensively to assist engineers with technical implementation, but for the first time AI systems can envision new systems, thus enabling higher-level cognitive tasks,” said Francesco Stella, lead author of the case study. can automate.” “This may involve a change of human roles to more technical ones.”

Based on technical tips provided by ChatGPT-3, the researchers built their robotic gripper and tested it in the real world, using it to pick tomatoes, which it did successfully.

Stella et al./EPFL/TU Delft

The researchers say that their case study demonstrates the potential to transform the design process through collaboration between humans and LLMs, but they are aware that this opens the door for varying degrees of collaboration.

At one extreme, he says, AI will act as an ‘inventor’, providing the entirety of robot design input and humans blindly applying it. An alternative would be to use AI’s extensive knowledge to supplement human expertise. A third approach would be to retain the human as an inventor and use AI to refine the design process through troubleshooting, debugging, and handling tedious or time-consuming processes.

The researchers weigh in on the ethical and common sense risks that may arise from human-AI collaboration. They point to issues of bias, plagiarism, and intellectual property (IP) rights as areas of concern and question whether the LLM-produced design can be considered ‘novel’, given that it Uses existing knowledge.

“In our study, ChatGPT identified tomato as the ‘highest value’ crop for robotic harvesters,” Hughes said. “However, it may be biased towards crops that are more covered in the literature, as opposed to those where there is actually a real need. When decisions are made outside the engineer’s realm of knowledge, this can lead to significant ethical, engineering or factual There may be errors.”

Despite these concerns, the researchers believe that human-AI collaboration has great potential if managed well.

“The robotics community must identify how to leverage these powerful tools to accelerate the advancement of robots in an ethical, sustainable and socially empowering manner,” the researchers said. “Looking ahead, we firmly believe that the LLM will open up many exciting possibilities and that if opportunity is managed, they will become a force for good.”

The case study was published in the journal nature machine intelligence,

Source: EPFL