Risk Assessor: Oversight of Physical AI Dangers

Audrey Miller, a partner at venture capital firm Tapestry VC, has voiced concerns that the current AI safety debate is overly focused on controlling large language models (LLMs), neglecting the significant risks associated with physical AI.

Miller notes that a growing number of projects are dedicated to improving the safety of physical AI systems, an area she describes as rapidly becoming a "massive category." She emphasizes that physical AI systems, which interact directly with the real world such as in robotics or autonomous vehicles, can cause tangible harm through accidents or malfunctions.

"The conversation is too much focused on AI safety, but it hasn't really tackled physical AI yet," Miller stated in an interview. "We need to think about how we can ensure these systems are safe and reliable."

Reports indicate that investments in AI within Europe have surged. In 2023, AI companies operating in Europe raised €28.5 billion in funding, more than double the amount raised in 2022. Despite this growth, the identification and management of physical AI risks require further attention.

Miller suggests that investors must recognize the potential of physical AI as a "massive category" that demands specific focus on safety considerations. She anticipates seeing more investments and innovations targeting these crucial safety challenges within physical AI.