Positronic robots bound by the Three Laws
Humanoid robots with general intelligence whose behavior is hard-wired to never harm humans, to obey orders, and to protect themselves, in that priority order.
In shortPositronic robots bound by the Three Laws, from I, Robot (Isaac Asimov, 1950), is at technology readiness level 3 of 9 (experimental proof of concept) at its weakest link. Held back by Artificial general intelligence, Continual learning without catastrophic forgetting, Formal verification of neural network behavior, Mechanistic interpretability of neural networks and Verifiable AI alignment. Science status: underlying science well understood; an engineering problem. Estimated new funding to unlock: $640B–3.7T.
Stories such as "Runaround" and "Liar!" follow US Robots' positronic robots and the paradoxes that arise when the Three Laws conflict.
Closest todayCommercial humanoids (Agility Digit, 1X NEO) and constitution-trained AI models; neither offers guaranteed behavioral rules.
Dependency stack
- Robots reason, converse and act across open-ended human situations
- Reliability collapses on tasks longer than a few hours at high success thresholds
- No persistent learning from experience after deployment
- Weak grounding in physical and social context outside text and code
TRL 3- Must keep learning from experience like a human employee
- Weight updates on new data degrade prior tasks and safety alignment (spurious forgetting)
- No method to validate a model whose parameters keep changing after release
- Per-user weight updates multiply serving cost and memory
TRL 3- Online updates and replay at frontier scale need large training compute in the loop
✕2 blockers›
- Grid power and interconnection queues
- Advanced-node chip and HBM memory supply
TRL 8
- Base reasoning, language and tool-use competence
✕2 blockers›
- Reliability and calibration over long horizons
- Continual learning after deployment
TRL 7- Training and inference computeTRL 8
- Training and running agents at human-equivalent throughputTRL 8
- The Laws must be guaranteed constraints, not learned tendencies
- Verification cost scales exponentially with network size for complete methods
- Behavioral rules like "do not harm a human" have no formal specification over raw inputs
- Properties proven on bounded input sets do not cover open-world inputs
TRL 3- Decomposing a model into understood parts enables compositional proofs
- Explanations cover only a fraction of model computation
TRL 3
- Interpreting "harm" correctly in novel situations is a value-specification problem
- No formal account of what to verify for open-ended agents
- Evaluations can be gamed by capable systems
TRL 3- Provable behavioral constraints on learned systemsTRL 3
- Inspecting internal goals rather than only behaviorTRL 3
- Robots perform arbitrary physical tasks
✕2 blockers›
- Robot training data at scale
- Tactile sensing and robust hands
TRL 5- Fine force control and slip detection need touch
✕3 blockers›
- Elastomer skins wear out and need recalibration after thousands of contacts
- Wiring and bandwidth for thousands of taxels across a whole hand
- Few large tactile datasets to train manipulation policies
TRL 5 - Generalist perception and planningTRL 7
- Human-form bodies working in human spaces
✕3 blockers›
- Cooperative safety certification for robots working in contact with people
- Battery life of 2–4 hours versus 8+ hour shifts
- Task generality beyond tote handling and pick-and-place
TRL 6- Hands and policies to handle arbitrary objectsTRL 5
- Touch feedback for safe contact and fine manipulationTRL 5
- Full-shift untethered operation
✕2 blockers›
- Cycle life of lithium-metal anodes
- Thermal runaway safety
TRL 6
- supportingArtificial muscle actuatorsHumanlike strength and quiet compliance in humanoid models
✕3 blockers›
- Kilovolt supply electronics and dielectric breakdown risk
- Low force per actuator and limited cycle life versus electric motors
- No integrated sensing comparable to muscle spindles
TRL 4
Who could unlock it
- Across several requirements:
- Anthropic 4
- METR 2
- Chris Olah 2
- OpenAIStartupBuilding Stargate with Oracle, SoftBank and MGX, with ~7 GW planned by Sept 2025 toward 10 GW by 2028; states AGI as its mission
- Google DeepMindCompanyReleased Gemini 3 (Dec 2025); Demis Hassabis frames AGI as the lab's goal
- AnthropicStartupClaude models led METR long-task time-horizon measurements in 2025-2026
- Demis HassabisResearcher, Google DeepMindGoogle DeepMind CEO and 2024 Nobel laureate (AlphaFold); directs Gemini and the lab's AGI research
- Ali BehrouzResearcher, Google Research (and Cornell University)First author of Titans (2025) and Nested Learning / Hope (NeurIPS 2025), the self-modifying architecture with a continuum memory system that reduced forgetting on continual-learning benchmarks
- Google ResearchCompanyPublished the Nested Learning paradigm (2025) and the follow-on "sleep" consolidation work (2026) that distils short-term into long-term memory
- Bing LiuResearcher, University of Illinois ChicagoLong-running lifelong-learning programme; his in-context continual learner with an external learner (InCA, 2025) is the standard non-parametric baseline
- TrajectoryStartupRaised a $15M seed in 2026 and then a Sequoia-led round to build a platform that keeps agentic models training on signals from live products
- Huan ZhangResearcher, University of Illinois Urbana-ChampaignLeads alpha-beta-CROWN, which won VNN-COMP every year from 2021 through 2025 and defines the state of the art in bound-propagation verification
- alpha-beta-CROWN (Verified Intelligence)Academic labOpen-source verifier that topped all scored benchmarks at VNN-COMP 2025; proves robustness properties for networks up to a few million parameters
- International Verification of Neural Networks Competition (VNN-COMP)ProgramIts sixth edition (2025) ran 8 teams over 16 regular and 9 extended benchmarks, and is the field's only standardised measure of what is provable at what scale
- Clark BarrettResearcher, Stanford UniversityCo-created Marabou, the SMT-based verifier used where bound propagation is too loose, and co-organises the benchmark standardisation effort
- Chris OlahResearcher, AnthropicFounded the circuits agenda; co-led Scaling Monosemanticity (2024) and the attribution-graph work (2025) that traced multi-step reasoning inside production models
- AnthropicCompanyOpen-sourced circuit-tracing and attribution-graph tooling in May 2025 with Neuronpedia, and has made interpretability an explicit corporate priority
- Neel NandaResearcher, Google DeepMindLeads DeepMind's mechanistic interpretability team and shipped Gemma Scope (2024), the open sparse-autoencoder suite most external work builds on
- GoodfireStartupRaised a $50M Series A in April 2025 (Menlo Ventures, with Anthropic participating) to turn interpretability into commercial model-auditing and steering tools
- AnthropicStartupRuns large interpretability (circuit tracing, 2025) and alignment-stress-testing teams and publishes model-organism studies of deceptive behavior
- Chris OlahResearcher, AnthropicPioneer of mechanistic interpretability; leads Anthropic's work mapping features and circuits inside production models (2024-2025)
- Redwood ResearchAcademic labNonprofit that founded the AI control research agenda (2023-2025), testing protocols that stay safe even if a model is scheming
- Buck ShlegerisResearcher, Redwood ResearchCEO of Redwood Research and lead author of AI control work; founding partner of the UK Alignment Project
- Physical IntelligenceStartupBuilds pi0-series vision-language-action robot policies (pi*0.6, Nov 2025); raised ~$600M at a $5.6B valuation in Nov 2025
- Sergey LevineResearcher, UC Berkeley / Physical IntelligenceCo-founder of Physical Intelligence; leads robot learning research on generalist manipulation policies
- Google DeepMind Gemini RoboticsCompanyReleased Gemini Robotics and Gemini Robotics-ER (2025), vision-language-action models built on Gemini 2.0 that do dexterous tasks such as origami folding
- Toyota Research InstituteCompanyDevelops large behavior models based on diffusion policy and has taught 60+ dexterous skills from demonstrations (2023-2025)
- Agility RoboticsStartupDigit humanoid in paid warehouse deployments (GXO and others); raised a $400M round in 2025
- Figure AIStartupRaised $1B+ Series C at a $39B valuation (Sept 2025) to scale Figure humanoids with in-house Helix AI
- ApptronikStartupApollo humanoid; Series A grew to $935M+ with a $520M extension in Feb 2026
- Tesla OptimusCompanyTesla's in-house humanoid program, aiming at factory use and mass production (2025-2026)
Beyond engineering
- Liability when a robot's rule interpretation causes harm
- Labor displacement and public fear of robots (the "Frankenstein complex")