Engineers at Princeton University and Google have come up with a new way to teach robots to know when they don’t know. The technique involves quantifying the fuzziness of human language and using that measurement to tell robots when to ask for further directions. Telling a robot to pick up a bowl from a table with only one bowl is fairly clear. But telling a robot to pick up a bowl when there are five bowls on the table generates a much higher degree of uncertainty — and triggers the robot to ask for clarification.

Because tasks are typically more complex than a simple “pick up a bowl” command, the engineers use large language models (LLMs) — the technology behind tools such as ChatGPT — to gauge uncertainty in complex environments. LLMs are bringing robots powerful capabilities to follow human language, but LLM outputs are still frequently unreliable, said Anirudha Majumdar, an assistant professor of mechanical and aerospace engineering at Princeton and the senior author of a study outlining the new method.

“Blindly following plans generated by an LLM could cause robots to act in an unsafe or untrustworthy manner, and so we need our LLM-based robots to know when they don’t know,” said Majumdar.

What happens when AI surpasses human-level intelligence? And WHEN exactly is this likely to happen?

That’s the focus of the next Metatrend in this Age of Abundance series.

Human-level AI, often referred to as AGI (artificial general intelligence) or ASI (artificial super intelligence), has historically been defined as the ability of a machine program to pass the “Turing Test,” defined as the ability of an AI to perform human-level tasks in a fashion indistinguishable from us humans. This definition is no longer useful.

The Taiwan Semiconductor Manufacturing Company’s (TSMC) advanced product costs could soar and hamper growth in A.I. chip supply.

Philospher of A.I.

The question of whether artificial intelligence could ever achieve consciousness is a common theme in science fiction. Could robots ever truly feel anything—like love, hate, or fear—or would they be all “dark inside”, experiencing nothing at all?

It is more important than ever to answer this question correctly. Artificial intelligence (AI) is no longer merely a matter of science fiction. AI are increasingly capable of producing art and mastering the use of language, raising serious questions about whether AI are already capable of consciousness, or if not yet, then soon.

How do we ensure that our robots are the most reliable on the market, capable of walking greater distances, climbing more steps, and inspecting more assets than any other mobile robot? In this webinar, we’ll offer an unprecedented glimpse into the rigorous testing each robot endures before it leaves our lab. We’ll also explore our commitment to customer success, covering our after-sales service and support programs that enable an exceptional customer experience.

In this webinar you will learn: