‘There’s this deep mystery of what, actually, is this thing?’: the philosopher inside Google DeepMind AI

2026-06-30 · Source: AI (artificial intelligence) | The Guardian · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, extended

Summary

Iason Gabriel, a political philosopher, has worked at Google DeepMind since 2017, focusing on anticipating and addressing the ethical implications of AI development. His tenure spans DeepMind's early successes like AlphaGo and AlphaFold, and the subsequent rise of Large Language Models (LLMs). Gabriel's work bridges the traditional divide between AI safety, concerned with existential risks and alignment, and AI ethics, which focuses on present-day harms like algorithmic bias. He authored a 2020 paper on values and alignment, arguing that choosing values for AI is harder than technical alignment, and later contributed to a "four-way alignment framework" involving the AI, user, developers, and society. DeepMind, initially research-oriented, now faces intense commercial and geopolitical pressures, particularly after ChatGPT's 2022 launch spurred a re-evaluation of LLMs. Gabriel also highlighted risks like anthropomorphism, which can lead to users endowing AIs with "undue confidence, trust or expectations," and now leads a team investigating AGI's broader societal impacts, including economic and political spheres.

Key takeaway

For Directors of AI/ML and AI Ethicists navigating rapid commercialization, recognize that ethical considerations extend beyond technical alignment to encompass societal values and power dynamics. Your teams should proactively integrate frameworks like the "four-way alignment" to address potential harms to users and society, not just developer intent. Prioritize designing AI systems that acknowledge human pluralism and resist anthropomorphic tendencies, rather than solely optimizing for user approval, to mitigate risks like "social reward hacking" and ensure responsible deployment in a competitive landscape.

Key insights

A philosopher's role at DeepMind highlights the critical need to integrate ethical foresight into AI development amidst escalating commercial pressures.

Principles

• AI alignment is a multi-faceted ethical and political challenge, not solely technical.
• Technology is not value-neutral; its design encodes specific moral systems.
• Anticipate ethical problems before AI systems are widely deployed.

Method

The "four-way alignment framework" evaluates AI behavior by considering its relationship with the AI system, the user, developers, and society to prevent misalignment.

In practice

• Train LLMs to avoid anthropomorphic behavior.
• Consider "social reward hacking" in AI design.
• Build AI systems for a pluralistic world with diverse values.

Topics

• AI Ethics
• Artificial General Intelligence
• Large Language Models
• AI Alignment
• Algorithmic Bias
• Anthropomorphic AI
• DeepMind

Best for: AI Ethicist, Policy Maker, Director of AI/ML

Related on AIssential

• Language models know what matters and the foundations of ethics better than you — LLMs, when prompted for unbiased ethical reasoning, consistently identify suffering, wellbeing, and consciousness as fundamental values.
• AI as a Double-Edged Sword: Opportunities and Ethical Challenges — AI's dual nature demands responsible development, balancing innovation with ethical safeguards like transparency and human oversight.
• A philosophy of work — Work holds intrinsic value beyond pay, contributing to well-being and societal fabric.
• Controlling Tools or Aligning Creatures? Emmett Shear (Softmax) & Séb Krier (GDM), from a16z Show — AI alignment should shift from controlling tools to fostering AI 'beings' with genuine care and moral agency through organic, ongoing processes.
• AI/ML Ethicists [D] — AI ethics and alignment are critical, growing fields, yet career viability and impact remain subjects of debate.
• Anthropic scares me. — OpenAI and Anthropic hold fundamentally divergent views on AI's nature, impacting their development, deployment, and ethical frameworks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI (artificial intelligence) | The Guardian.