Sponsored Session: Who Do the Robots Work For? The Surprisingly Simple Answer to AI's... Dave Lockie

# Overview In this talk, Dave Lockie, a strategy wrangler at Automattic, addresses AI alignment and modern technological development from a human and embodied perspective. Prompted by Tim Berners-Lee's question, "Who does the AI work for?", Lockie explores how AI systems and robots reflect the values, states, and intentions of those who create them. He argues that technological alignment is not solely a technical problem solvable through nonprofits, open source, or blockchain consensus mechanisms, but is deeply intertwined with human emotional states, burnout, and embodied cognition.

# Key Topics - **The Alignment Question:** Lockie discusses how robots and AI embody different values depending on their creators—ranging from commercial vacuum cleaners to autonomous vehicles and Buddhist temple robots. He points out that AI proliferates widely and is trained on collective human digital exhaust. - **Limitations of Conventional Tech Solutions:** In seeking alignment mechanisms, Lockie examined nonprofits (referencing OpenAI's shift toward corporate structures), open-source software (which can also be used maliciously and suffers from maintainer burnout), and crypto/Web3 systems (noting Bitcoin's consensus is narrowly focused on ledger states rather than complex human values). - **The Hyper Ego and Cognitive Overload:** Lockie describes how immersive screen use disconnects people from physical reality, trapping them in rational mental loops driven by the personal super-ego. When connected across global digital networks, this aggregates into what he terms the "hyper ego"—a disembodied, reactive collective intelligence that AI models ingest and mirror. - **Embodied Intelligence and the Gibbons:** Citing an interview by James Bridle, Lockie shares an evolutionary biology study on tool use among primates. While other great apes easily used tools on the ground to retrieve food, gibbons failed until tools were hung overhead. Because gibbons brachiate in the tree canopy, their intelligence is embodied and adapted to upward movement. Lockie uses this to demonstrate that intelligence cannot be separated from physical embodiment and environmental context.

# Technical Details - **Audience Slido Experiment:** Lockie conducted an interactive experiment using Slido to test word association tasks under two physiological states: first while overstimulated and distracted by phones/notifications (leading to fragmented, scattered responses), and second after guided triangle breathing (inhale 3s, exhale 3s, hold empty 3s), illustrating how collective performance and coherence change with biological state. - **Information Networks (Truth vs. Order):** Referencing Yuval Noah Harari's book *Nexus*, Lockie highlights the tension in information networks between generating truth versus producing order and power, noting corporate AI models often prioritize predictable profits, order, and engagement over wisdom and human well-being. - **Reframing Techniques:** Drawing inspiration from Laurie Voss (creator of npm), Lockie suggests cognitive reframings—such as viewing AI as "additional intelligence" rather than "artificial intelligence," and reframing problems as puzzles.

# Takeaways - **"Robots Work for Whoever Builds Them":** The ultimate answer to who robots work for is simple: they serve the creators who build them. The true challenge lies in deciding who we want to be and what values we bring while building these systems. - **Embodied Self-Regulation:** Tech workers and open-source maintainers facing burnout should integrate practices to drop out of purely rational overdrive and reconnect with their bodies (e.g., triangle breathing, movement, breaks). - **Inefficiency as a Feature:** Embracing deliberate slowness, nuance, and space for human connection prevents systems from encoding tyranny and prioritizing speed and efficiency over humanity.