Google introduced a set of rules called the “Robot Constitution” on Thursday. These guidelines will oversee the artificial intelligence (AI) controlling their upcoming smart machines.
Google’s robots follow specific guidelines inspired by Isaac Asimov’s Three Laws of Robotics. These rules make sure the robots don’t hurt people and avoid tasks involving sharp objects or electrical tools. Yet, Google knows these rules alone aren’t enough to keep everyone safe. They also use other safety methods like controlling how much force the robots use and having a human supervisor who can shut them down if needed.
To achieve this, Google’s data collection system called AutoRT uses two types of models: a visual language model (VLM) and a large language model (LLM). As an extra safety measure, DeepMind programmed the robots to stop by themselves if too much pressure is put on their joints. They also added a physical kill switch that human operators can use to turn them off. Over seven months, Google sent out 53 AutoRT robots to four office buildings. They ran more than 77,000 tests. Some robots were guided by people from a distance, some followed a set plan, and others worked independently using Google’s Robotic Transformer (RT-2) AI learning model.
Also, SARA-RT, or self-adaptive robust attention for robotic transformers, is a fresh system designed to make the decision-making in AI transformer models quicker.
What are your thoughts on this News?
Comment Your Thoughts Below.
Visit https://ainews.instamart.ai/ to read interesting AI News Daily
Follow our Instagram for Daily AI Content: Click here