Figure AI establishes center for the safety of humanoid robots

The robotics company Figure AI, developer of the humanoid robot Figure 02, which has already been sold to companies and is working with humans in factories on a trial basis, is setting up a center to promote the safety of humanoid robots. This was announced on LinkedIn on Tuesday by Rob Gruendel, Chief Safety Engineer at Figure AI. The aim is to develop safety standards for humanoid robots that must be met before they are allowed to work together with humans.

Dangerous robots

In some cases, humanoid robots are already working together with humans in factories. The robots are used because they can be easily integrated into work environments made for humans and can work there without having to make changes to the environment. In such environments, collaboration with humans should also be largely seamless.

However, this is currently more of a hope than a reality, as humanoid robots can potentially injure their human colleagues if they malfunction. There are no safety mechanisms, such as shutdown devices like the familiar emergency stop button, for humanoid robots. This violates safety standards. There are also no guidelines to prevent collisions between humans and humanoid robots.

Development of safety standards

However, safe cooperation is the be-all and end-all when it comes to convincing people to work hand in hand with humanoid robots. This is why Figure AI is in the process of setting up a Center for the Advancement of Humanoid Safety. The center will evaluate and improve the safety of the company's own humanoid robots. To this end, standards and guidelines are to be developed to enable safe cooperation with humans. To this end, the engineers want to test the stability of AI-controlled humanoid robots when stationary and in motion. In addition, the recognition of humans and animals, safe AI behaviors and evasive movements to avoid injuries are to be tested and improved, writes Gruendel.

Figure AI is hoping for further input from customers and the public on how to increase the safety of humanoid robots. Ultimately, safety standards for humanoid robots are to be developed based on all the information collected. Figure AI also wants to advocate for robots to be certified depending on the application. For example, it is not enough to test a “safety device for presence detection according to a laboratory or office IT standard if it is used in a harsh industrial environment or outdoors,” Gruendel explains in the announcement. He therefore advocates harmonized testing standards and regulations, without which there can be no safe humanoid robots.

Figure AI plans to publish quarterly reports on updates and progress in the safety of humanoid robots. The robotics company promises that this will also include a list of failures.

(olb)