DeepMind Researcher Describes LLMs as ‘Exotic Mind-Like Entities’: The Need for New Terminology in AI’s Future
Understanding Artificial Intelligence: A Deeper Look
Artificial intelligence (AI) is revolutionizing our world, yet even top scientists admit they find it challenging to define exactly what AI is. Recently, a thought-provoking video clip shared on Instagram by the account theaigems featured Murray Shanahan, a Principal Scientist at DeepMind. In this clip, he described large language models (LLMs) as more than mere machines, calling them “exotic mind-like entities.”
The Challenge of Definition
Shanahan’s terminology raises interesting questions. Why does he use the term “mind-like”? Current language does not adequately capture the essence of these new digital intelligences. While they exhibit language skills similar to humans, they lack physical forms, resulting in a unique notion of selfhood that feels unfamiliar. The hyphen in "mind-like" serves as a cautious bridge, suggesting these entities mirror some aspects of human minds but do not possess consciousness in a traditional sense.
DeepMind: A Leader in AI Research
Murray Shanahan’s insights gain more significance when considering his role at DeepMind Technologies, one of the leading AI research labs, which was established in 2010 and is now a subsidiary of Alphabet Inc. DeepMind has made remarkable strides in artificial intelligence, showcasing its capabilities in notable projects.
- AlphaGo: DeepMind made headlines when its AI defeated a world champion in the complex board game Go.
- AlphaFold: This project has dramatically advanced the prediction of protein structures, transforming the field of biology.
In 2023, DeepMind merged with Google Brain to create Google DeepMind, solidifying its position as a global leader in AI research.
The Mind Behind AI: Murray Shanahan
As a professor of Cognitive Robotics at Imperial College London and a senior scientist at DeepMind, Murray Shanahan has extensively explored the intersection of AI, cognitive science, and philosophy. His expertise is further underscored by his role as a scientific advisor for the 2014 film Ex Machina, where he guided the discussion surrounding the nature of AI consciousness.
Shanahan has consistently voiced concerns about the difficulties in defining intelligence for entities that do not resemble human beings. His description of LLMs as “exotic” reflects a comprehensive understanding of the complexities surrounding these evolving systems. As AI continues to develop at a rapid pace, societal frameworks struggle to keep up with their comprehension.
A New Framework for Understanding AI
Shanahan’s statement sheds light on a critical necessity: We must develop new ways to think about intelligence, selfhood, and consciousness as AI systems become more integrated into our daily lives. This emerging reality could require us to re-evaluate fundamental concepts long associated with humanity.
- Intelligence: How do we define intelligence in beings that may not have human-like experiences?
- Selfhood: What does it mean to have a sense of self in a digital context?
- Consciousness: Can we assign consciousness to something created from code and algorithms?
As we admire the capabilities of AIs developed by institutions like DeepMind, we must acknowledge an unsettling reality — we are crafting new kinds of minds that challenge our understanding. In our quest to explore these cutting-edge technologies, humanity is poised at a philosophical crossroad, still in need of the right terminology to describe the entities born from advanced programming and complex algorithms.
Exploring these ideas could help society develop a more nuanced understanding of the evolving relationship between humans and AI, opening discussions that bridge technology and philosophy.
