Pith. sign in

REVIEW 3 major objections 4 minor 6 cited by

Existing robot, vehicle, and virtual-agent policies leave most embodied-AI risks under-governed, and the gaps demand urgent but targeted fixes.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

A policy analysis arguing that embodied AI risks are real, under-covered by current US/EU/UK frameworks, and best handled through certification, benchmarks, clarified liability, and economic adaptation.

T0 review reviewed 2026-08-05 challenge →

load-bearing objection A coherent policy synthesis on embodied AI governance with a useful risk taxonomy and a Table 1 that overclaims precision, but it deserves serious review. the 3 major comments →

arxiv 2509.00117 v2 pith:TQJ5MGMY submitted 2025-08-28 cs.CY cs.AIcs.RO

Embodied AI: Emerging Risks and Opportunities for Policy Action

Jared Perlo , Alexander Robey , Fazl Barez , Luciano Floridi , Jakob M\"okander This is my paper
classification cs.CY cs.AIcs.RO
keywords embodied AIAI governancerisk taxonomyrobotics policyautonomous vehiclesAI safetyliabilitylabor displacement
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

Embodied AI—agents that learn, reason, and act in the physical world—is advancing faster than the rules meant to govern it. The paper argues that existing policies for industrial robots, autonomous vehicles, and virtual agentic AI do not cover the full set of risks embodied systems create, and that the gaps are urgent enough for policymakers to act now. It builds its case in three steps: a taxonomy of physical, informational, economic, and social harms; a coverage analysis of US, EU, and UK frameworks; and a set of targeted recommendations, including mandatory certification, clearer liability, post-deployment monitoring, and economic adaptation. If the argument holds, the path forward is not sweeping new legislation but deliberate extension and patching of existing frameworks before a crisis forces rushed rulemaking.

Core claim

The paper's central claim is that embodied AI systems—defined as the intersection of agentic AI and classical robots—pose a distinct bundle of risks that current governance neither fully covers nor coherently assigns to a regulator. It identifies four risk families: physical (malicious and accidental harm), informational (privacy and misinformation), economic (labor displacement, inequality, power concentration), and social (bias, accountability, transparency, unhealthy attachment, transformative effects). It then rates existing US, EU, and UK policies against these risks and finds that while some frameworks—the EU Machinery Regulation, the EU AI Act, the GDPR, drone rules, and the UK's Auto

What carries the argument

The analytical engine is the four-part risk taxonomy—physical, informational, economic, and social—each subdivided into concrete subrisks, paired with a coverage matrix (Table 1) that rates existing policies for classical robots, autonomous vehicles, and virtual agents on each subrisk using a three-level scale (substantial, partial, or little coverage). The taxonomy organizes the field; the matrix turns that organization into a gap analysis. On top sits a recommendation stack—certification, standards, liability, monitoring, and economic adaptation—that is meant to attach to the gaps the matrix reveals.

Load-bearing premise

The policy-gap conclusion rests on the assumption that the hand-selected US, EU, and UK statutes and the subjective coverage ratings in Table 1 accurately represent how well existing governance covers every EAI risk.

What would settle it

A re-rating of Table 1 by independent policy analysts using a published coding protocol could settle the claim: if most risk subcategories were rated substantially covered by existing statutes, the paper's conclusion of critical gaps would collapse. Alternatively, a documented case where a current statute already mandates pre-deployment certification and post-market monitoring for a general-purpose mobile EAI system would undercut the claim that no such framework exists.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

If this is right

  • Mandatory pre-deployment testing and certification for EAI systems, tiered by operational context and capability, becomes a concrete near-term policy target rather than an abstract call for oversight.
  • Liability frameworks need to designate a responsible entity for fully autonomous operation, extending precedents such as the UK's Authorized Self-Driving Entity to other mobile EAI forms.
  • Post-deployment monitoring and black-box-style data retention rules need to be harmonized across the EU Machinery Regulation and the AI Act, whose current one-year versus six-month retention requirements conflict.
  • Economic and social risks—labor displacement, inequality, power concentration, and human-EAI attachment—require proactive policy blueprints because they arise from EAI working too well, not from technical failure.
  • Industry-led standards can move faster than legislation and can fill near-term gaps, but voluntary standards alone will not cover the statutory holes in certification and enforcement.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the Table 1 ratings are even roughly right, the first enforceable EAI rules will likely come from extending familiar AV-style operational-design-domain certification to other mobile robots, because that is where regulatory precedents already exist.
  • The paper's four-part taxonomy could be operationalized as a risk registry: each identified subrisk can be mapped to concrete evaluable metrics, turning a policy argument into an engineering target for benchmark developers.
  • The internal contradiction between the Machinery Regulation's one-year data-retention mandate and the AI Act's six-month minimum suggests that harmonization of existing rules, rather than new law, may be the quickest concrete safety win.
  • If the observation that 'embodied AI' is partly a marketing label is taken seriously, regulators should audit whether new products actually fall under existing machinery, drone, or vehicle rules before creating new regimes.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper argues that embodied AI (EAI) systems present physical, informational, economic, and social risks that are not adequately covered by existing US/EU/UK frameworks governing industrial robots, autonomous vehicles, and virtual agents. It contributes (1) a four-part risk taxonomy, (2) a qualitative policy-gap analysis summarized in Table 1, and (3) policy recommendations including certification, benchmarks, liability clarification, post-deployment monitoring, and economic adaptation measures. The central claim is that existing policies are insufficient and should be extended rather than replaced.

Significance. If taken as a policy-scoping document, the paper is useful and timely. Its risk taxonomy is broad and well referenced, and its recommendations are concrete and actionable. The paper is honest about several limitations (geographic scope, civilian focus, overlap with LLM risks) and engages with counterarguments such as market forces and hardware constraints. The main contribution is qualitative synthesis rather than formal analysis; there are no fitted parameters, machine-checked proofs, or empirical datasets. The credibility of the policy-gap conclusion therefore rests heavily on the transparency and rigor of the coverage assessment in Table 1, which currently lacks auditability.

major comments (3)
  1. [Section 3.2 / Table 1] The load-bearing claim that existing governance has 'significant and concerning gaps' is operationalized through the H/partial/# ratings in Table 1, but the table is presented without a coding protocol, a source matrix linking each cell to specific statutory provisions, or inter-rater reliability checks. A different analyst could reasonably rate, for example, the EU Machinery Regulation plus AI Act as providing substantial certification coverage for autonomous machinery, or GDPR Article 22 as an existing accountability mechanism, which would change several cells and weaken the conclusion. Section 5's limitations discuss geography and application scope but not this methodology. The authors should either add a transparent coding procedure with a provision-by-provision source matrix, or explicitly reframe Table 1 as an illustrative expert judgment rather than an auditable evidence base.
  2. [Section 2.1, refs [66]-[68]] The claim that 'several recent reports document an increase in industrial injuries following the introduction of AI-controlled robots' is supported only by individual news stories about particular incidents, not by systematic occupational-injury data. This is not a fatal flaw, but the wording overstates the evidence. The authors should either hedge to 'documented incidents' or cite systematic studies of robot-related injury rates if available.
  3. [Section 3.2, data-retention example] The paper states that the EU Machinery Regulation's one-year data-retention requirement and the AI Act's 'at least six months' requirement are 'contradictory guidance.' This is not a contradiction: a one-year minimum satisfies a six-month minimum. If the authors intend to illustrate regulatory confusion, they should identify an actual conflict (e.g., different scopes, unclear interaction between MR and AI Act obligations) or remove this example.
minor comments (4)
  1. [Section 3.2] Typo: 'EAI sytems' should be 'EAI systems'.
  2. [Table 1 caption] The paper uses 'A Vs' with an awkward space in the table caption and elsewhere; consider 'AVs'.
  3. [Section 4.1] Typo/incomplete sentence: 'Researchers should and build on this progress' appears to be missing a verb.
  4. [Section 3.1] The discussion of GDPR Article 22 and autonomous EAI is useful but would benefit from a more precise statement of how Article 22's 'automated decision-making' test would apply to an embodied system that acts but does not necessarily make a 'decision' in the GDPR sense.

Circularity Check

0 steps flagged

No significant circularity: the policy analysis is self-contained; author self-citations are background support only.

full rationale

This is a policy analysis rather than a formal derivation: it constructs a risk taxonomy from external literature, surveys US/EU/UK statutes, assigns qualitative coverage ratings in Table 1, and proposes policy recommendations. There are no fitted parameters, no equations, and no 'prediction' that is statistically forced by construction. The central claim—that existing frameworks are insufficient—is made qualitatively and supported by the statutory survey (e.g., the EU Machinery Regulation/AI Act retention-period conflict, unclear liability for Level 4/5 AVs, and the absence of a post-deployment monitoring regime); it is not derived from a model fitted to the same conclusion. Table 1's coverage ratings are subjective expert codings without a protocol or inter-rater check, which is an evidence/reproducibility limitation, not circularity: the ratings are inputs to the gap analysis, and the paper does not claim they are outputs of an independent derivation. Several authors cite their own prior work (Mökander & Floridi on sociotechnical pragmatism and AI auditing; Robey on LLM jailbreaks; Barez et al. on AI-enabled authoritarianism), but these are used for background scope choices and risk plausibility, not as load-bearing justifications of the paper's main policy conclusion. No uniqueness theorem, ansatz, or fitted result is imported from the authors' own prior work. Accordingly, the derivation chain does not reduce to its own inputs.

Axiom & Free-Parameter Ledger

0 free parameters · 4 axioms · 0 invented entities

The paper is a qualitative policy analysis, so there are no fitted numbers or new postulated entities. Its central claim depends on assumptions about future EAI deployment, the distinctness of embodied risks, and the reliability of the authors' judgment-based coverage ratings in Table 1.

axioms (4)
  • domain assumption Embodied AI systems will be deployed at scale in civilian settings with increasing autonomy in the near future.
    The introduction and Section 2 repeatedly project that EAI capabilities and domains 'are likely to expand significantly in the coming years'; the urgency of the policy call rests on this trajectory.
  • domain assumption Physical embodiment creates risks materially distinct from virtual AI, warranting separate regulation.
    This is the central framing of the paper. The authors address overlap with LLM risks in Section 5 but maintain that embodiment makes risks immediate, pervasive, and scale-dependent.
  • ad hoc to paper The coverage ratings in Table 1 accurately reflect the adequacy of existing policies.
    No coding protocol, source matrix, or validation is given. The table is the authors' own judgment and directly drives the gap analysis in Section 3.2.
  • domain assumption The selected US, EU, and UK statutes and standards are representative of relevant governance for EAI.
    The limitations section acknowledges geography and scope choices, but no systematic selection criterion or search protocol is provided for which laws and standards were analyzed.

reviewed 2026-08-05 · how reviews work

0 comments
Cite this review

Pith. "Pith review of Embodied AI: Emerging Risks and Opportunities for Policy Action." pith.science (2026). https://pith.science/paper/TQJ5MGMY

@misc{pith2026250900117,
  author       = {Pith},
  title        = {Pith review of: Embodied AI: Emerging Risks and Opportunities for Policy Action},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TQJ5MGMY}},
  note         = {Machine review of arXiv:2509.00117}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

The field of embodied AI (EAI) is rapidly advancing. Unlike virtual AI, EAI systems can exist in, learn from, reason about, and act in the physical world. With recent advances in AI models and hardware, EAI systems are becoming increasingly capable across wider operational domains. While EAI systems can offer many benefits, they also pose significant risks, including physical harm from malicious use, mass surveillance, as well as economic and societal disruption. These risks require urgent attention from policymakers, as existing policies governing industrial robots and autonomous vehicles are insufficient to address the full range of concerns EAI systems present. To help address this issue, this paper makes three contributions. First, we provide a taxonomy of the physical, informational, economic, and social risks EAI systems pose. Second, we analyze policies in the US, EU, and UK to assess how existing frameworks address these risks and to identify critical gaps. We conclude by offering policy recommendations for the safe and beneficial deployment of EAI systems, such as mandatory testing and certification schemes, clarified liability frameworks, and strategies to manage EAI's potentially transformative economic and societal impacts.

Figures

Figures reproduced from arXiv: 2509.00117 by Alexander Robey, Fazl Barez, Jakob M\"okander, Jared Perlo, Luciano Floridi.

Figure 1
Figure 1. Figure 1: A comparison of classical robots, agentic AI, and EAI EAI represents the intersection of agentic AI and classical robots. Many existing robots, such as industrial machines (including articulated arms, gantry systems, and other types) lack the autonomy and reasoning capabilities that agentic AI possesses. Conversely, many agentic AI systems such as chatbots or virtual assistants currently lack physical embo… view at source ↗
Figure 2
Figure 2. Figure 2: An illustrated summary of risks from embodied AI. We identify four key risk categories and provide several existing or potential mechanisms through which EAI systems could cause harm within each risk area. spanning visual, auditory, and tactile information—during deployment [12]. Like text-based virtual AI models, which are known to memorize and expose personally identifiable information [75, 76], commerci… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 6 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Underwriting the Agent Economy: The Blueprint for an AI Insurance Stack

    cs.CY 2026-07 conditional novelty 6.5

    Affirmative AI-agent insurance with billion-scale limits is achievable by 2030 solely through coordinated industry build-out of an eight-component stack spanning data, CAT models, standards, contracts, underwriting, p...

  2. Underwriting the Agent Economy: The Blueprint for an AI Insurance Stack

    cs.CY 2026-07 conditional novelty 6.0

    A coordinated eight-component insurance infrastructure could make affirmative AI-agent coverage with billion-dollar limits achievable by 2030.

  3. Safety in Embodied AI: A Survey of Risks, Attacks, and Defenses

    cs.CR 2026-03 unverdicted novelty 6.0

    The survey organizes over 400 papers on embodied AI safety into a multi-level taxonomy and flags overlooked issues such as fragile multimodal fusion and unstable planning under jailbreaks.

  4. Safety in Embodied AI: A Survey of Risks, Attacks, and Defenses

    cs.CR 2026-03 accept novelty 6.0

    A multi-level taxonomy of risks, attacks, and defenses across the full embodied AI pipeline, synthesizing 500+ papers and flagging overlooked failure modes.

  5. Physical AI Governance: From Theory to Practice Across Life Cycle

    cs.AI 2026-07 conditional novelty 4.0

    A survey that organizes Physical AI governance into five principles and a five-stage lifecycle, with stage-specific operational practices.

  6. A Co-Evolutionary Theory of Human-AI Coexistence: Mutualism, Governance, and Dynamics in Complex Societies

    cs.CY 2026-04 unverdicted novelty 4.0

    Human-AI coexistence is best modeled as conditional mutualism under governance, formalized as a multiplex dynamical system whose simulations show stable high-coexistence equilibria only under balanced institutional oversight.

Reference graph

Works this paper leans on

218 extracted references · 36 canonical work pages · cited by 4 Pith papers · 3 internal anchors

  1. [1]

    A call for embodied AI, September 2024

    Giuseppe Paolo, Jonas Gonzalez-Billandon, and Balázs Kégl. A call for embodied AI, September 2024. URL http://arxiv.org/abs/2402.03824. arXiv:2402.03824 [cs]

    Pith/arXiv arXiv 2024

  2. [2]

    Aligning Cyber Space with Physical World: A Comprehensive Survey on Embodied AI, August 2024

    Yang Liu, Weixing Chen, Yongjie Bai, Xiaodan Liang, Guanbin Li, Wen Gao, and Liang Lin. Aligning Cyber Space with Physical World: A Comprehensive Survey on Embodied AI, August 2024. URL http://arxiv.org/abs/2407.06886. arXiv:2407.06886 [cs]

    Pith/arXiv arXiv 2024

  3. [3]

    Drone- Aided Delivery Methods, Challenge, and the Future: A Methodological Review

    Xueping Li, Jose Tupayachi, Aliza Sharmin, and Madelaine Martinez Ferguson. Drone- Aided Delivery Methods, Challenge, and the Future: A Methodological Review. Drones, 7(3):191, March 2023. ISSN 2504-446X. doi: 10.3390/drones7030191. URL https: //www.mdpi.com/2504-446X/7/3/191

    work page doi:10.3390/drones7030191 2023

  4. [4]

    Chinese researchers develop amphibious spherical robot assisting in po- lice patrol in E.China’s Zhejiang

    Du Qiongfang. Chinese researchers develop amphibious spherical robot assisting in po- lice patrol in E.China’s Zhejiang. Global Times, December 2024. URL https://www. globaltimes.cn/page/202412/1324773.shtml

    2024

  5. [5]

    AI robots may hold key to nursing Japan’s ageing population

    Kiyoshi Takenaka. AI robots may hold key to nursing Japan’s ageing population. Reuters, February 2025. URL https: //www.reuters.com/technology/artificial-intelligence/ ai-robots-may-hold-key-nursing-japans-ageing-population-2025-02-28/

    2025

  6. [6]

    AI robots are helping South Korea’s seniors feel less alone

    Michelle Kim. AI robots are helping South Korea’s seniors feel less alone. Rest of World , August 2025. URL https://restofworld.org/2025/ korea-ai-robot-senior-care-hyodol/?utm_source=linkedin&utm_medium= social&utm_campaign=row-social

    2025

  7. [7]

    The rise of ai robots: Physical ai is coming for you

    Rob Garlick, Wenyan Fei, Tahmid Quddus Islam, Anjola Odunsi, Adam Spielman, Martin Wilkie, Helen Krause, Matthew Moffat, Carol Gibson, Alex Miller, and Anuj Gangahar. The rise of ai robots: Physical ai is coming for you. Technical report, Citi GPS: Global Perspectives & Solutions, December 2024. URL https://ir. citi.com/gps/H558-XNr_iTlGa7Qq7H9AYb5ZT2W851...

    2024

  8. [8]

    The coming wave: Technology, Power, and the Twenty-First Century’s Greatest Dilemma

    Mustafa Suleyman and Michael Bhaskar. The coming wave: Technology, Power, and the Twenty-First Century’s Greatest Dilemma. Crown, New York, 2023. ISBN 978-0-593-59395-0 978-0-593-72817-8

    2023

  9. [9]

    Atkinson

    Robert D. Atkinson. Robots and International Economic Development. ITIF Policy Report, January 2021. URL ssrn.com/abstract=3875581

    2021

  10. [10]

    Prescott and Julie M

    Tony J. Prescott and Julie M. Robillard. Are friends electric? The benefits and risks of human-robot relationships. iScience, 24(1):101993, January 2021. ISSN 25890042. doi: 10.1016/j.isci.2020.101993. URL https://linkinghub.elsevier.com/retrieve/pii/ S2589004220311901

    arXiv 2021

  11. [11]

    The Future of Embodied AI: Containing and Mitigating the Dark and Creepy Sides of Robotics, Autonomous Vehicles, and AI

    Jo Ann Oravec. The Future of Embodied AI: Containing and Mitigating the Dark and Creepy Sides of Robotics, Autonomous Vehicles, and AI. In Good Robot, Bad Robot, pages 245–276. Springer International Publishing, Cham, 2022. ISBN 978-3-031-14012-9 978-3-031-14013-

    2022

  12. [12]

    URL https://link.springer.com/10.1007/ 978-3-031-14013-6_9

    doi: 10.1007/978-3-031-14013-6_9. URL https://link.springer.com/10.1007/ 978-3-031-14013-6_9 . Series Title: Social and Cultural Studies of Robots and AI

    work page doi:10.1007/978-3-031-14013-6_9

  13. [13]

    OpenVLA: An Open-Source Vision-Language-Action Model, September 2024

    Moo Jin Kim, Karl Pertsch, Siddharth Karamcheti, Ted Xiao, Ashwin Balakrishna, Suraj Nair, Rafael Rafailov, Ethan Foster, Grace Lam, Pannag Sanketi, Quan Vuong, Thomas Kollar, Benjamin Burchfiel, Russ Tedrake, Dorsa Sadigh, Sergey Levine, Percy Liang, and Chelsea Finn. OpenVLA: An Open-Source Vision-Language-Action Model, September 2024. URL http://arxiv....

    Pith/arXiv arXiv 2024

  14. [14]

    All feedback is welcome

    Gemini Robotics Team, Saminda Abeyruwan, Joshua Ainslie, Jean-Baptiste Alayrac, Montser- rat Gonzalez Arenas, Travis Armstrong, Ashwin Balakrishna, Robert Baruch, Maria Bauza, Michiel Blokzijl, Steven Bohez, Konstantinos Bousmalis, Anthony Brohan, Thomas Buschmann, Arunkumar Byravan, Serkan Cabi, Ken Caluwaerts, Federico Casarini, Oscar Preprint. All feed...

    Pith/arXiv arXiv 2025

  15. [15]

    Qwen2.5-VL Techni- cal Report, February 2025

    Shuai Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, Sibo Song, Kai Dang, Peng Wang, Shijie Wang, Jun Tang, Humen Zhong, Yuanzhi Zhu, Mingkun Yang, Zhaohai Li, Jian- qiang Wan, Pengfei Wang, Wei Ding, Zheren Fu, Yiheng Xu, Jiabo Ye, Xi Zhang, Tianbao Xie, Zesen Cheng, Hang Zhang, Zhibo Yang, Haiyang Xu, and Junyang Lin. Qwen2.5-VL Techni- cal Repor...

    Pith/arXiv arXiv 2025

  16. [16]

    URL https://nvidianews.nvidia.com/news/ nvidia-isaac-gr00t-n1-open-humanoid-robot-foundation-model-simulation-frameworks

    NVIDIA Announces Isaac GR00T N1 — the World’s First Open Humanoid Robot Foundation Model — and Simulation Frameworks to Speed Robot De- velopment, March 2025. URL https://nvidianews.nvidia.com/news/ nvidia-isaac-gr00t-n1-open-humanoid-robot-foundation-model-simulation-frameworks

    2025

  17. [17]

    China pits humanoid robots against humans in half-marathon for first time

    Alessandro Diviggiano and Eduardo Baptista. China pits humanoid robots against humans in half-marathon for first time. Reuters, April 2025. URL https://www.reuters.com/world/china/ china-pits-humanoid-robots-against-humans-half-marathon-2025-04-19/

    2025

  18. [18]

    URL https: //www.figure.ai/news/scaling-helix-logistics

    Scaling Helix: a New State of the Art in Humanoid Logistics, June 2025. URL https: //www.figure.ai/news/scaling-helix-logistics

    2025

  19. [19]

    Technical report, Physical Intelligence, February 2025

    Open sourcing pi0. Technical report, Physical Intelligence, February 2025. URL https: //www.physicalintelligence.company/blog/openpi

    2025

  20. [20]

    URL https:// huggingface.co/unitreerobotics

    Unitree open-source embodied intelligence datasets and models. URL https:// huggingface.co/unitreerobotics

  21. [21]

    data gap

    Ken Goldberg. Good old-fashioned engineering can close the 100,000-year “data gap” in robotics. Science Robotics, 10(105):eaea7390, August 2025. ISSN 2470-9476. doi: 10.1126/ scirobotics.aea7390. URL https://www.science.org/doi/10.1126/scirobotics. aea7390

    doi:10.1126/scirobotics 2025

  22. [22]

    All feedback is welcome

    Open X.-Embodiment Collaboration, Abby O’Neill, Abdul Rehman, Abhinav Gupta, Abhiram Maddukuri, Abhishek Gupta, Abhishek Padalkar, Abraham Lee, Acorn Pooley, Agrim Gupta, Ajay Mandlekar, Ajinkya Jain, Albert Tung, Alex Bewley, Alex Herzog, Alex Irpan, Alexander Khazatsky, Anant Rai, Anchit Gupta, Andrew Wang, Andrey Kolobov, Anikait Singh, Animesh Garg, A...

  23. [23]

    Recent Advances in Tactile Sensory Systems: Mechanisms, Fabrication, and Applications

    Jianguo Xi, Huaiwen Yang, Xinyu Li, Ruilai Wei, Taiping Zhang, Lin Dong, Zhenjun Yang, Zuqing Yuan, Junlu Sun, and Qilin Hua. Recent Advances in Tactile Sensory Systems: Mechanisms, Fabrication, and Applications. Nanomaterials, 14(5):465, March 2024. ISSN 2079-4991. doi: 10.3390/nano14050465. URL https://www.mdpi.com/2079-4991/14/ 5/465

    work page doi:10.3390/nano14050465 2024

  24. [24]

    Robot Dexterity – Handling our future

    Jenny Read. Robot Dexterity – Handling our future. Technical report, Advanced Re- search and Invention Agency, 2024. URL https://www.aria.org.uk/media/xamlbwdo/ aria-robotic-dexterity-programme-thesis.pdf

    2024

  25. [25]

    Data will solve robotics and automation: True or false?

    Nancy M. Amato, Seth Hutchinson, Animesh Garg, Aude Billard, Daniela Rus, Russ Tedrake, Frank Park, and Ken Goldberg. “Data will solve robotics and automation: True or false?”: A debate. Science Robotics , 10(105):eaea7897, August 2025. ISSN 2470-

    2025

  26. [26]

    Real-Time Execution of Action Chunk- ing Flow Policies, June 2025

    Kevin Black, Manuel Galliker, and Sergey Levine. Real-Time Execution of Action Chunk- ing Flow Policies, June 2025. URL https://www.physicalintelligence.company/ download/real_time_chunking.pdf

    2025

  27. [27]

    America Is Missing The New Labor Economy – Robotics Part 1

    Dylan Patel, Reyk Knuhtsen, Niko Ciminelli, Jeremie Eliahou Ontiveros, Joe Ryu, and Robert Ghilduta. America Is Missing The New Labor Economy – Robotics Part 1. Technical report, SemiAnalysis, March 2025. URL https://semianalysis.com/2025/03/ 11/america-is-missing-the-new-labor-economy-robotics-part-1/?access_ token#what-stands-to-come

    2025

  28. [28]

    Bloomberg, May 2025

    China’s Startups Race to Dominate the Coming AI Robot Boom. Bloomberg, May 2025. URL https://www.bloomberg.com/features/2025-china-ai-robots-boom/

    2025

  29. [29]

    Embodied AI Agents: Modeling the World, June 2025

    Pascale Fung, Yoram Bachrach, Asli Celikyilmaz, Kamalika Chaudhuri, Delong Chen, Willy Chung, Emmanuel Dupoux, Hervé Jégou, Alessandro Lazaric, Arjun Majumdar, Andrea Madotto, Franziska Meier, Florian Metze, Théo Moutakanni, Juan Pino, Basile Terver, Joseph Tighe, and Jitendra Malik. Embodied AI Agents: Modeling the World, June 2025. URL http://arxiv.org/...

    Pith/arXiv arXiv 2025

  30. [30]

    Luciano Floridi and J.W. Sanders. On the Morality of Artificial Agents. Minds and Ma- chines, 14(3):349–379, August 2004. ISSN 0924-6495, 1572-8641. doi: 10.1023/B: MIND.0000035461.63578.9d. URL https://link.springer.com/10.1023/B:MIND. 0000035461.63578.9d

    arXiv 2004

  31. [31]

    Characterizing AI Agents for Alignment and Gover- nance, 2025

    Atoosa Kasirzadeh and Iason Gabriel. Characterizing AI Agents for Alignment and Gover- nance, 2025. URL https://arxiv.org/abs/2504.21848. Version Number: 1

    Pith/arXiv arXiv 2025

  32. [32]

    Robots, Jobs, Taxes, and Responsibilities

    Luciano Floridi. Robots, Jobs, Taxes, and Responsibilities. Philosophy & Technology, 30 (1):1–4, March 2017. ISSN 2210-5433, 2210-5441. doi: 10.1007/s13347-017-0257-3. URL http://link.springer.com/10.1007/s13347-017-0257-3

    work page doi:10.1007/s13347-017-0257-3 2017

  33. [33]

    Human–robot collaboration and machine learning: A systematic review of recent research

    Francesco Semeraro, Alexander Griffiths, and Angelo Cangelosi. Human–robot collaboration and machine learning: A systematic review of recent research. Robotics and Computer- Integrated Manufacturing, 79:102432, 2023. ISSN 0736-5845. doi: https://doi.org/10.1016/ j.rcim.2022.102432. URL https://www.sciencedirect.com/science/article/pii/ S0736584522001156

    arXiv 2023

  34. [34]

    Mazumder, M.F

    A. Mazumder, M.F. Sahed, Z. Tasneem, P. Das, F.R. Badal, M.F. Ali, M.H. Ahamed, S.H. Abhi, S.K. Sarker, S.K. Das, M.M. Hasan, M.M. Islam, and M.R. Islam. Towards next generation digital twin in robotics: Trends, scopes, challenges, and future. Heliyon, 9(2): e13359, February 2023. ISSN 24058440. doi: 10.1016/j.heliyon.2023.e13359. URL https: //linkinghub....

    work page doi:10.1016/j.heliyon.2023.e13359 2023

  35. [35]

    Towards Robust and Secure Embodied AI: A Survey on Vulnerabilities and Attacks, February 2025

    Wenpeng Xing, Minghao Li, Mohan Li, and Meng Han. Towards Robust and Secure Embodied AI: A Survey on Vulnerabilities and Attacks, February 2025. URLhttp://arxiv.org/abs/ 2502.13175. arXiv:2502.13175 [cs]

    Pith/arXiv arXiv 2025

  36. [36]

    I, Robot

    Isaac Asimov. I, Robot. Panther, St. Albans, reprint edition, 1977. ISBN 978-0-586-02532-1

    1977

  37. [37]

    Cambridge University Press, Cambridge, United Kingdom New York, NY , 2024

    Woodrow Barfield, Yueh-Hsuan Weng, and Ugo Pagallo, editors.The Cambridge handbook on the law, policy, and regulation of human-robot interaction. Cambridge University Press, Cambridge, United Kingdom New York, NY , 2024. ISBN 978-1-00-938670-8

    2024

  38. [38]

    Autonomy 2.0: The Quest for Economies of Scale

    Shuang Wu, Bo Yu, Shaoshan Liu, and Yuhao Zhu. Autonomy 2.0: The Quest for Economies of Scale. Communications of the ACM, 68(4):28–32, April 2025. ISSN 0001-0782, 1557-7317. doi: 10.1145/3708012. URL https://dl.acm.org/doi/10.1145/3708012

    work page doi:10.1145/3708012 2025

  39. [39]

    Kemin Zhou, John Comstock Doyle, and John C. Doyle. Essentials of robust control. Prentice Hall international editions. Prentice Hall, Upper Saddle River, NJ, 1998. ISBN 978-0-13- 525833-0 978-0-13-790874-5. Preprint. All feedback is welcome. 19

    1998

  40. [40]

    Adaptive control

    Karl Johan Aström and Björn Wittenmark. Adaptive control. Addison-Wesley, Reading (Mass.), 2nd ed edition, 1995. ISBN 978-0-201-55866-1

    1995

  41. [41]

    A Taxonomy of Factors Influencing Perceived Safety in Human–Robot Interaction

    Neziha Akalin, Andrey Kiselev, Annica Kristoffersson, and Amy Loutfi. A Taxonomy of Factors Influencing Perceived Safety in Human–Robot Interaction. International Jour- nal of Social Robotics , 15(12):1993–2004, December 2023. ISSN 1875-4791, 1875-

    1993

  42. [42]

    Ames, Xiangru Xu, Jessy W

    Aaron D. Ames, Xiangru Xu, Jessy W. Grizzle, and Paulo Tabuada. Control Barrier Function Based Quadratic Programs for Safety Critical Systems. IEEE Transactions on Automatic Control, 62(8):3861–3876, August 2017. ISSN 0018-9286, 1558-2523. doi: 10.1109/TAC. 2016.2638961. URL http://ieeexplore.ieee.org/document/7782377/

    arXiv 2017

  43. [43]

    A Fully Automated Framework for Control of Linear Systems from Temporal Logic Specifications

    Marius Kloetzer and Calin Belta. A Fully Automated Framework for Control of Linear Systems from Temporal Logic Specifications. IEEE Transactions on Automatic Control , 53(1):287–297, February 2008. ISSN 0018-9286. doi: 10.1109/TAC.2007.914952. URL http://ieeexplore.ieee.org/document/4459804/

    arXiv 2008

  44. [44]

    Concrete Problems in AI Safety, July 2016

    Dario Amodei, Chris Olah, Jacob Steinhardt, Paul Christiano, John Schulman, and Dan Mané. Concrete Problems in AI Safety, July 2016. URL http://arxiv.org/abs/1606.06565. arXiv:1606.06565 [cs]

    Pith/arXiv arXiv 2016

  45. [45]

    Mayne, J.B

    D.Q. Mayne, J.B. Rawlings, C.V . Rao, and P.O.M. Scokaert. Constrained model pre- dictive control: Stability and optimality. Automatica, 36(6):789–814, June 2000. ISSN 00051098. doi: 10.1016/S0005-1098(99)00214-9. URL https://linkinghub.elsevier. com/retrieve/pii/S0005109899002149

    doi:10.1016/s0005-1098(99)00214-9 2000

  46. [46]

    Generating Robot Constitutions & Benchmarks for Semantic Safety, 2025

    Pierre Sermanet, Anirudha Majumdar, Alex Irpan, Dmitry Kalashnikov, and Vikas Sindhwani. Generating Robot Constitutions & Benchmarks for Semantic Safety, 2025. URL https: //arxiv.org/abs/2503.08663. Version Number: 1

    Pith/arXiv arXiv 2025

  47. [47]

    URL https://docs.un.org/en/A/C.1/79/L.77

    General and complete disarmament: lethal autonomous weapons systems, December 2024. URL https://docs.un.org/en/A/C.1/79/L.77

    2024

  48. [48]

    Michael J. D. Vermeer, Tim Bonds, Emily Lathrop, and Gregory Smith. Averting a Robot Catastrophe: Preparing for Converging Trends in Robotics and Frontier AI, April 2025. URL https://osf.io/ymvf5_v1

    2025

  49. [49]

    Cooperative inverse reinforcement learning

    Dylan Hadfield-Menell, Stuart J Russell, Pieter Abbeel, and Anca Dragan. Cooperative inverse reinforcement learning. In D. Lee, M. Sugiyama, U. Luxburg, I. Guyon, and R. Gar- nett, editors, Advances in Neural Information Processing Systems , volume 29. Curran As- sociates, Inc., 2016. URL https://proceedings.neurips.cc/paper_files/paper/ 2016/file/c3395dd...

    2016

  50. [50]

    Watson, Jakob Mökander, and Luciano Floridi

    David S. Watson, Jakob Mökander, and Luciano Floridi. Competing narratives in AI ethics: a defense of sociotechnical pragmatism. AI & SOCIETY, December 2024. ISSN 0951-5666, 1435-5655. doi: 10.1007/s00146-024-02128-2. URL https://link.springer.com/10. 1007/s00146-024-02128-2

    work page doi:10.1007/s00146-024-02128-2 2024

  51. [51]

    Auditing large language models: a three-layered approach

    Jakob Mökander, Jonas Schuett, Hannah Rose Kirk, and Luciano Floridi. Auditing large language models: a three-layered approach. AI and Ethics , 4(4):1085–1115, November

  52. [52]

    As ai advances, u.s

    Eric Schmitt and Charlie Savage. As ai advances, u.s. seeks to keep autonomous weapons in check. The New York Times, November 2023. URL https://www.nytimes.com/2023/11/ 21/us/politics/ai-drones-war-law.html . Preprint. All feedback is welcome. 20

    2023

  53. [53]

    An Approach to Technical AGI Safety and Security, April 2025

    Rohin Shah, Alex Irpan, Alexander Matt Turner, Anna Wang, Arthur Conmy, David Lindner, Jonah Brown-Cohen, Lewis Ho, Neel Nanda, Raluca Ada Popa, Rishub Jain, Rory Greig, Samuel Albanie, Scott Emmons, Sebastian Farquhar, Sébastien Krier, Senthooran Rajamanoha- ran, Sophie Bridgers, Tobi Ijitoye, Tom Everitt, Victoria Krakovna, Vikrant Varma, Vladimir Mikul...

    Pith/arXiv arXiv 2025

  54. [54]

    Exclusive: Silicon valley startup breaks cover with plans for robo-armies

    Colin Demarest. Exclusive: Silicon valley startup breaks cover with plans for robo-armies. Axios, April 2025. URL https://www.axios.com/2025/04/16/ scout-ai-military-autonomous-fury

    2025

  55. [55]

    What we know about ukraine’s army of robot dogs

    David Hambling. What we know about ukraine’s army of robot dogs. Forbes, August 2024. URL https://www.forbes.com/sites/davidhambling/2024/08/16/ what-we-know-about-ukraines-army-of-robot-dogs/

    2024

  56. [56]

    Cummings

    M.L. Cummings. Artificial Intelligence and the Future of Warfare. Technical report, Chatham House, January 2017. URL https://www. chathamhouse.org/sites/default/files/publications/research/ 2017-01-26-artificial-intelligence-future-warfare-cummings-final.pdf

    2017

  57. [57]

    Universal and transferable adversarial attacks on aligned language models

    Andy Zou, Zifan Wang, Nicholas Carlini, Milad Nasr, J Zico Kolter, and Matt Fredrikson. Universal and transferable adversarial attacks on aligned language models. arXiv preprint arXiv:2307.15043, 2023

    Pith/arXiv arXiv 2023

  58. [58]

    Ukraine’s future vision and current capabilities for wag- ing ai-enabled autonomous warfare

    Kateryna Bondar. Ukraine’s future vision and current capabilities for wag- ing ai-enabled autonomous warfare. Center for Strategic and International Studies (CSIS) , March 2025. URL https://www.csis.org/analysis/ ukraines-future-vision-and-current-capabilities-waging-ai-enabled-autonomous-warfare

    2025

  59. [59]

    Jailbreakbench: An open robustness benchmark for jailbreaking large language models

    Patrick Chao, Edoardo Debenedetti, Alexander Robey, Maksym Andriushchenko, Francesco Croce, Vikash Sehwag, Edgar Dobriban, Nicolas Flammarion, George J Pappas, Florian Tramer, et al. Jailbreakbench: An open robustness benchmark for jailbreaking large language models. arXiv preprint arXiv:2404.01318, 2024

    Pith/arXiv arXiv 2024

  60. [60]

    Smoothllm: Defending large language models against jailbreaking attacks

    Alexander Robey, Eric Wong, Hamed Hassani, and George J Pappas. Smoothllm: Defending large language models against jailbreaking attacks. arXiv preprint arXiv:2310.03684, 2023

    Pith/arXiv arXiv 2023

  61. [61]

    Jailbreaking llm-controlled robots

    Alexander Robey, Zachary Ravichandran, Vijay Kumar, Hamed Hassani, and George J Pappas. Jailbreaking llm-controlled robots. In 2025 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2025

    2025

  62. [62]

    Badrobot: Manipulating embodied llms in the physical world

    Hangtao Zhang, Chenyu Zhu, Xianlong Wang, Ziqi Zhou, Changgan Yin, Minghui Li, Lulu Xue, Yichen Wang, Shengshan Hu, Aishan Liu, et al. Badrobot: Manipulating embodied llms in the physical world. arXiv preprint arXiv:2407.20242, 2024

    Pith/arXiv arXiv 2024

  63. [63]

    Jailbreaking black box large language models in twenty queries

    Patrick Chao, Alexander Robey, Edgar Dobriban, Hamed Hassani, George J Pappas, and Eric Wong. Jailbreaking black box large language models in twenty queries. arXiv preprint arXiv:2310.08419, 2023

    Pith/arXiv arXiv 2023

  64. [64]

    Adversarial attacks on robotic vision language action models

    Eliot Krzysztof Jones, Alexander Robey, Andy Zou, Zachary Ravichandran, George J Pappas, Hamed Hassani, Matt Fredrikson, and J Zico Kolter. Adversarial attacks on robotic vision language action models. arXiv preprint arXiv:2506.03350, 2025

    Pith/arXiv arXiv 2025

  65. [65]

    Exploring the adversarial vulnerabilities of vision-language-action models in robotics

    Taowen Wang, Cheng Han, James Chenhao Liang, Wenhao Yang, Dongfang Liu, Luna Xinyu Zhang, Qifan Wang, Jiebo Luo, and Ruixiang Tang. Exploring the adversarial vulnerabilities of vision-language-action models in robotics. arXiv preprint arXiv:2411.13587, 2024

    Pith/arXiv arXiv 2024

  66. [66]

    Man crushed to death by robot in south korea, November 2023

    Emily Atkinson. Man crushed to death by robot in south korea, November 2023. URL https://www.bbc.com/news/world-asia-67354709 . Accessed: 2025-05-12

    2023

  67. [67]

    Tesla robot attacks worker in austin factory and leaves them bleeding, December 2023

    Fortune Staff. Tesla robot attacks worker in austin factory and leaves them bleeding, December 2023. URL https://fortune.com/2023/12/27/ tesla-factory-robot-worker-attack-injury/ . Accessed: 2025-05-12

    2023

  68. [68]

    Safety guardrails for llm-enabled robots

    Zachary Ravichandran, Alexander Robey, Vijay Kumar, George J Pappas, and Hamed Hassani. Safety guardrails for llm-enabled robots. arXiv preprint arXiv:2503.07885, 2025

    arXiv 2025

  69. [69]

    AI Alignment: A Comprehensive Survey, April 2025

    Jiaming Ji, Tianyi Qiu, Boyuan Chen, Borong Zhang, Hantao Lou, Kaile Wang, Yawen Duan, Zhonghao He, Lukas Vierling, Donghai Hong, Jiayi Zhou, Zhaowei Zhang, Fanzhi Zeng, Juntao Dai, Xuehai Pan, Kwan Yee Ng, Aidan O’Gara, Hua Xu, Brian Tse, Jie Fu, Stephen McAleer, Yaodong Yang, Yizhou Wang, Song-Chun Zhu, Yike Guo, and Wen Gao. AI Alignment: A Comprehensi...

    Pith/arXiv arXiv 2025

  70. [70]

    Human–robot collaboration in manufacturing applications: a review

    Eloise Matheson, Riccardo Minto, Emanuele GG Zampieri, Maurizio Faccio, and Giulio Rosati. Human–robot collaboration in manufacturing applications: a review. Robotics, 8(4): 100, 2019

    2019

  71. [71]

    Redefining Robotics with Carolina Parada

    Hannah Fry. Redefining Robotics with Carolina Parada. URL https://podcasts.apple. com/ro/podcast/redefining-robotics-with-carolina-parada/id1476316441? i=1000709450547

  72. [72]

    Alan F. T. Winfield, Matimba Swana, Jonathan Ives, and Sabine Hauert. On the ethical governance of swarm robotic systems in the real world. Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences , 383(2289):20240142, January 2025. ISSN 1364-503X, 1471-2962. doi: 10.1098/rsta.2024.0142. URL https: //royalsocietyp...

    arXiv 2025

  73. [73]

    24 amazon workers sent to hospital after robot accidentally un- leashes bear spray, December 2018

    Soo Youn. 24 amazon workers sent to hospital after robot accidentally un- leashes bear spray, December 2018. URL https://abcnews.go.com/US/ 24-amazon-workers-hospital-bear-repellent-accident/story?id=59625712 . Accessed: 2025-05-12. Preprint. All feedback is welcome. 21

    2018

  74. [74]

    SafeAPT: Safe Simulation-to-Real Robot Learning using Diverse Policies Learned in Simulation

    Rituraj Kaushik, Karol Arndt, and Ville Kyrki. SafeAPT: Safe Simulation-to-Real Robot Learning using Diverse Policies Learned in Simulation, 2022. URL https://arxiv.org/ abs/2201.13248. Version Number: 1

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  75. [75]

    Scalable extraction of training data from (production) language models

    Milad Nasr, Nicholas Carlini, Jonathan Hayase, Matthew Jagielski, A Feder Cooper, Daphne Ippolito, Christopher A Choquette-Choo, Eric Wallace, Florian Tramèr, and Katherine Lee. Scalable extraction of training data from (production) language models. arXiv preprint arXiv:2311.17035, 2023

    Pith/arXiv arXiv 2023

  76. [76]

    The secret sharer: Evaluating and testing unintended memorization in neural networks

    Nicholas Carlini, Chang Liu, Úlfar Erlingsson, Jernej Kos, and Dawn Song. The secret sharer: Evaluating and testing unintended memorization in neural networks. In 28th USENIX security symposium (USENIX security 19), pages 267–284, 2019

    2019

  77. [77]

    Ryan Calo

    M. Ryan Calo. The Boundaries of Privacy Harm. Indiana Law Journal, 86:1131–1162, 2011. URL https://www.repository.law.indiana.edu/ilj/vol86/iss3/8/

    2011

  78. [78]

    Using simulation and domain adaptation to improve efficiency of deep robotic grasping

    Konstantinos Bousmalis, Alex Irpan, Paul Wohlhart, Yunfei Bai, Matthew Kelcey, Mrinal Kalakrishnan, Laura Downs, Julian Ibarz, Peter Pastor, Kurt Konolige, Sergey Levine, and Vincent Vanhoucke. Using simulation and domain adaptation to improve efficiency of deep robotic grasping. In 2018 IEEE International Conference on Robotics and Automation (ICRA), pag...

    arXiv 2018

  79. [79]

    A Roomba recorded a woman on the toilet

    Eileen Guo. A Roomba recorded a woman on the toilet. How did screenshots end up on Facebook? MIT Technology Review , December

  80. [80]

    AI-Enabled Coups: How a Small Group Could Use AI to Seize Power

    Tom Davidson, Lukas Finnveden, and Rose Hadshar. AI-Enabled Coups: How a Small Group Could Use AI to Seize Power. Technical report, Fore- thought Institute, April 2025. URL https://www.forethought.org/research/ ai-enabled-coups-how-a-small-group-could-use-ai-to-seize-power

    2025

Showing first 80 references.

This paper was first reviewed by deepseek-v4-flash on August 5, 2026.