Google turns to sci-fi for input on ‘robot constitution’
Breakthroughs in robotics could have a significant impact on both work environments and broader industrial applications, making human-robot interaction more efficient and safer. However as robots are integrated into human workloads the temptation exists to use them in situations that could put humans at harm.
Now the robotics team at Google DeepMind (owned by Alphabet) has announced three innovations that enable robots to make faster, better and safer decisions in practical situations. One is a ‘constitution for robots’.
Taking its inpiration from the work of science fiction writer Isaac Asimov, the constition consists of three basic principles:
- A robot must not injure a human being or by doing nothing allow a human being to be injured
- A robot must obey orders given to it by humans unless such orders violate the first law
- A robot must protect its own existence as long as such protection does not violate first or second law
DeepMind is also introducing an advanced data collection system, AutoRT, that combines a visual language model (VLM) and a large language model (LLM). The company said in a blog post that over the course of seven months, 53 AutoRT robots were tested in different office environments, with more than 77,000 trials. The robots could understand their environment and choose tasks using VLM and LLM.
These technologies are part of a broader trend toward more autonomous and intelligent robotic systems, whose applications range from office assistance to more complex industrial tasks.
DeepMind’s latest developments in robotics promise a future where robots can work seamlessly and safely with humans, marking a major step forward in the practical application of robotics technology.
News Wires
Subscribers 0
Fans 0
Followers 0
Followers