pith. sign in

arxiv: 2604.18380 · v1 · submitted 2026-04-20 · 💻 cs.AI · cs.CY

The implicated scientist: on the role of AI researchers in the development of weapons systems

Alexandra Volokhova , Alex Hernandez-Garcia This is my paper

Pith reviewed 2026-05-10 04:56 UTC · model grok-4.3

classification 💻 cs.AI cs.CY
keywords AI ethicsweapons systemsmoral implicationAI arms racesolidarityimplicated subjectstechnological harms
0
0 comments X

The pith

AI researchers are morally implicated subjects in the harms caused by AI-enabled weapons systems and can transform this into solidarity with victims.

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

The paper examines the increasing use of AI in modern weapons systems that have been linked to mass killings and an accelerating arms race among powerful nations. It positions AI researchers as implicated in these harms due to their contributions to the underlying technologies. The authors explore how this implication can be redirected toward differentiated, long-distance solidarity with those affected by the resulting injustices and inequalities. Readers would care because it highlights the ethical stakes for scientists in a field driving both technological progress and potential large-scale destruction.

Core claim

AI researchers function as implicated subjects in the development of weapons systems that enable mass killings and increased inequality through the AI arms race. By investigating the specifics of this implication, the work proposes transfiguring the researchers' position into one of differentiated, long-distance solidarity with the victims of these technologically fortified injustices.

What carries the argument

The position of the implicated AI researcher in weapons development, which frames their technical contributions as creating moral responsibility and enables a shift to solidarity with victims.

If this is right

  • Recognizing implication prompts researchers to consider their role in preventing further harms from AI weapons.
  • The ongoing AI arms race extends implication to more researchers and heightens risks of devastation and inequality.
  • Transfiguring implication into solidarity can address technologically fortified injustices directly.
  • This approach differentiates responsibilities based on specific involvement in AI tech for military purposes.

Where Pith is reading between the lines

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

  • The concept could extend to researchers contributing to other high-risk technologies with dual civilian-military uses.
  • Actions like public advocacy or policy engagement might follow from adopting solidarity as a response.
  • It raises questions about how academic institutions should handle funding from defense sectors.
  • Broader ethical training in AI could incorporate discussions of implication and solidarity.

Load-bearing premise

That the work of AI researchers creates a chain of moral implication connecting them individually to the harms inflicted by weapons systems using AI.

What would settle it

A demonstration that AI research contributions are sufficiently isolated from weapons deployment decisions or that researchers lack any awareness or intent regarding military applications would falsify the direct implication claim.

read the original abstract

Artificial intelligence (AI) technologies are increasingly used in modern weapons systems. Notably, these systems have recently been involved in mass killings and destruction at scale. Furthermore, there is currently a strong interest and competition among powerful players to accelerate the proliferation of weapons with automated or AI-based components, a phenomenon known as AI arms race. This competition poses a risk of causing even more deaths and devastation in the future, as well as increased power and wealth inequality. In this work, we aim to shed light on the role of AI researchers as implicated subjects in the harms caused by weapons enabled by AI technologies. We investigate and discuss the specifics of this implication and explore ways to transfigure this position of implication into one of differentiated, long-distance solidarity with the victims of technologically fortified injustices.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The paper argues that AI researchers are 'implicated subjects' in harms caused by AI-enabled weapons systems amid an AI arms race, investigates the specifics of this moral implication, and proposes ways to transfigure the position into differentiated long-distance solidarity with victims of technologically fortified injustices.

Significance. If the interpretive argument holds, the work contributes to AI ethics by framing researcher responsibility in military applications and suggesting pathways for professional solidarity, which could inform debates on dual-use technologies and arms control.

major comments (1)
  1. Abstract: The central claim that AI researchers are directly implicated in harms from weapons systems does not specify how this applies across roles, institutions, or levels of involvement (e.g., academic theorists vs. those contributing code to defense contractors), which is load-bearing for the scope of the implication argument and the subsequent transfiguration proposal.
minor comments (1)
  1. The abstract and title use 'implicated scientist' and 'implicated subjects' without an early definition or reference to prior literature on implication in ethics, which would aid clarity for a cs.AI readership.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We appreciate the referee's constructive feedback, which has helped us refine the precision of our central argument. We address the major comment below and have incorporated revisions to strengthen the manuscript.

read point-by-point responses

  1. Referee: Abstract: The central claim that AI researchers are directly implicated in harms from weapons systems does not specify how this applies across roles, institutions, or levels of involvement (e.g., academic theorists vs. those contributing code to defense contractors), which is load-bearing for the scope of the implication argument and the subsequent transfiguration proposal.

    Authors: We thank the referee for identifying this point of clarification. The abstract employs a broad framing to reflect the structural nature of implication in an AI arms race, where contributions range from foundational academic research with clear dual-use implications to direct technical work for defense entities. The full manuscript elaborates on these variations in the sections discussing implicated subjects and pathways to solidarity. To ensure the claim's scope is explicit and supports the transfiguration proposal, we have revised the abstract to note that the implication applies to AI researchers across academic, industry, and institutional roles whose work contributes to the development or acceleration of AI-enabled weapons systems. This change improves precision without narrowing the paper's overall thesis. revision: yes

Circularity Check

0 steps flagged

No significant circularity in interpretive ethical analysis

full rationale

The paper is a philosophical and normative discussion examining the ethical position of AI researchers in relation to AI-enabled weapons systems. It advances an interpretive argument about implication and solidarity without any equations, derivations, empirical fits, parameter estimations, or technical predictions. No load-bearing step reduces by construction to its own inputs, self-citations, or renamed known results. The central claim rests on conceptual framing and ethical reasoning rather than deductive or statistical chains that could be shown equivalent to prior assumptions within the paper's own structure. This is a self-contained interpretive work whose premises are not derived from fitted data or self-referential definitions.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on unproven ethical assumptions about researcher implication and responsibility that are introduced as domain premises rather than derived from evidence or prior technical results.

axioms (2)
  • domain assumption AI researchers are implicated subjects in harms from AI-enabled weapons systems
    This is the load-bearing premise invoked to frame the discussion of role and solidarity.
  • domain assumption An AI arms race exists and increases future risks of death and inequality
    Stated as background fact without supporting data or citations in the provided abstract.

pith-pipeline@v0.9.0 · 5428 in / 1385 out tokens · 47624 ms · 2026-05-10T04:56:17.318261+00:00 · methodology

discussion (0)

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

Reference graph

Works this paper leans on

30 extracted references · 30 canonical work pages

  1. [1]

    The grey hoodie project: Big tobacco, big tech, and the threat on academic integrity

    Mohamed Abdalla and Moustafa Abdalla. The grey hoodie project: Big tobacco, big tech, and the threat on academic integrity. InProceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, pp. 287–297,

    work page 2021

  2. [2]

    ‘Lavender’: The AI machine directing Israel’s bombing spree in Gaza, 2024a

    Yuval Abraham. ‘Lavender’: The AI machine directing Israel’s bombing spree in Gaza, 2024a. URL https://www.972mag.com/lavender-ai-israeli-army-gaza . Accessed: 2025-08-25. Yuval Abraham. ‘Order from Amazon’: How tech giants are storing mass data for Israel’s war, 2024b. URL https://www.972mag.com/cloud-israeli-army-gaza-amazon-googl e-microsoft. Accessed:...

    work page 2025

  3. [3]

    Accessed: 2026-03-19

    URL https://www.972mag.com/microsoft-8200-intelligence-surveillanc e-cloud-azure. Accessed: 2026-03-19. Nur Ahmed, Muntasir Wahed, and Neil C Thompson. The growing influence of industry in AI research.Science, 379(6635):884–886,

    work page 2026

  4. [4]

    Accessed: 2025-03-07

    URL https://theintercept.com/2024/01/12/open-ai-military-ban-chatgpt . Accessed: 2025-03-07. 6 Yasmin Astrid Bonea. Algorithms of war: AI, drones and the challenge to international humanitarian law in the Russia-Ukraine war.Faculty of Law, Lund University,

    work page 2024

  5. [5]

    Accessed: 2026-02-04

    URL https://www.canada.ca/en/depar tment-national-defence/programs/borealis.html. Accessed: 2026-02-04. Stephanie Russo Carroll, Ibrahim Garba, Oscar L Figueroa-Rodr´ıguez, Jarita Holbrook, Raymond Lovett, Simeon Materechera, Mark Parsons, Kay Raseroka, Desi Rodriguez-Lonebear, Robyn Rowe, et al. The CARE principles for Indigenous data governance.Open Sch...

    work page 2026

  6. [6]

    Andrea Coveri, Claudio Cozza, and Dario Guarascio

    URL https://www.europarl.e uropa.eu/thinktank/en/document/EPRS_BRI(2025)769580. Andrea Coveri, Claudio Cozza, and Dario Guarascio. Blurring boundaries: an analysis of the digital platforms-military nexus.Review of Political Economy, 37(4):1632–1663,

    work page 2025

  7. [7]

    Microsoft blocks Israel’s use of its technology in mass surveillance of Palestinians, 2025a

    Harry Davies and Yuval Abraham. Microsoft blocks Israel’s use of its technology in mass surveillance of Palestinians, 2025a. URL https://www.theguardian.com/world/2025/sep/2 5/microsoft-blocks-israels-use-of-its-technology-in-mass-surve illance-of-palestinians. Accessed: 2026-03-19. Harry Davies and Yuval Abraham. Revealed: Microsoft deepened ties with Is...

    work page 2025

  8. [8]

    Accessed: 2026-02-05

    URL https://techworkerhandbook.org . Accessed: 2026-02-05. Hamid Ekbia. Killing in the name of precision.Science for the people, vol. 25, no.3, Killing in the Name Of,

    work page 2026

  9. [9]

    Accessed: 2026-02-04

    URL https://www.consilium.euro pa.eu/en/policies/defence-numbers. Accessed: 2026-02-04. Marwa Fatafta and Daniel Leufer. Artificial genocidal intelligence: How Israel is automating human rights abuses and war crimes,

    work page 2026

  10. [10]

    Accessed: 2025-08-28

    URL https://www.accessnow.org/publicatio n/artificial-genocidal-intelligence-israel-gaza. Accessed: 2025-08-28. Luca Franceschi, Michele Donini, C´edric Archambeau, and Matthias Seeger. Explaining probabilistic models with distributional values.arXiv preprint arXiv:2402.09947,

    work page arXiv 2025

  11. [11]

    Accessed: 2026-01-28

    URL https: //www.forbes.com/sites/allbusiness/2026/01/07/compensation-for-a i-employees-is-skyrocketing. Accessed: 2026-01-28. Alex Hernandez-Garcia, Alexandra V olokhova, Ezekiel Williams, and Dounia Shaaban Kabakibo. Irresponsible AI: big tech’s influence on AI research and associated impacts.arXiv preprint arXiv:2512.03077,

    work page arXiv 2026

  12. [12]

    Accessed: 2025-08-27

    URL https://www.icj-cij.org/case/192. Accessed: 2025-08-27. Amnesty International. Automated apartheid: How facial recognition fragments, segregates and controls Palestinians in the OPT,

    work page 2025

  13. [13]

    Accessed: 2025-03-07

    URL https://www.amnesty.org/en/documen ts/mde15/6701/2023/en. Accessed: 2025-03-07. Pratyusha Ria Kalluri, William Agnew, Myra Cheng, Kentrell Owens, Luca Soldaini, and Abeba Birhane. The surveillance AI pipeline.arXiv preprint arXiv:2309.15084,

    work page arXiv 2023

  14. [14]

    Accessed: 2026-02-05

    URL https://medium.com/@ arielkoren/googles-complicity-in-israeli-apartheid-how-google-w eaponizes-diversity-to-silence-palestinians-cb41b24ac423 . Accessed: 2026-02-05. Aarathi Krishnan, Angie Abdilla, AJung Moon, Carlos Affonso Souza, Chelle Adamson, Eileen M. Lach, Farah Ghazal, Jessica Fjeld, Jennyfer Taylor, John C. Havens, Malavika Jayaram, Monique ...

    work page 2026

  15. [15]

    Accessed: 2026-02-04

    URL https: //manyfesto.ai. Accessed: 2026-02-04. Jacob Larson, James S. Denford, Gregory S. Dawson, and Kevin C. Desouza. The evolution of artificial intelligence (AI) spending by the U.S. government,

    work page 2026

  16. [16]

    Accessed: 2026-02-03

    URL https://www.broo kings.edu/articles/the-evolution-of-artificial-intelligence-ai-s pending-by-the-u-s-government. Accessed: 2026-02-03. Lockheed Martin. Lockheed Martin and Google Cloud collaborate to advance generative AI for national security,

    work page 2026

  17. [17]

    Accessed: 2025-08-29

    URL https://www.lockheedmartin.com/en-us/news/fea tures/2025/lockheed-martin-google-collaborate-to-advance-generat ive-ai-national-security.html. Accessed: 2025-08-29. Lucas Maaser and Stephanie Verlaan. Big tech goes to war: Uncovering the growing role of US and European technology firms in the military–industrial complex.Studien, 5:2022,

    work page 2025

  18. [18]

    Accessed: 2026-02-04

    URL LockheedMartin,Microsoftanno uncelandmarkagreementonclassifiedcloud,advancedtechnologiesforDe partmentofDefense. Accessed: 2026-02-04. NAFA. No Azure for Apartheid Petition,

    work page 2026

  19. [19]

    Accessed: 2026-03-19

    URL https://noazureforapartheid.co m/petition. Accessed: 2026-03-19. NATO. Defence expenditures and NATO’s 5 % commitment,

    work page 2026

  20. [20]

    Accessed: 2026-02-04

    URLhttps://www.nato .int/en/what-we-do/introduction-to-nato/defence-expenditures-and -natos-5-commitment. Accessed: 2026-02-04. Gerrit De Vynck Nitasha Tiku. Google drops pledge not to use AI for weapons or surveillance,

    work page 2026

  21. [21]

    Accessed: 2025-03-07

    URLhttps://archive.is/nNNlD. Accessed: 2025-03-07. 8 NTFA. No tech for apartheid, 2026a. URL https://www.notechforapartheid.com . Accessed: 2026-02-05. NTFA. Tell Google: Stop using your AI for genocide, apartheid, and border violence, 2026b. URL https://actionnetwork.org/letters/tell-google-stop-using-your-a i-for-genocide-apartheid-and-border-violence. ...

    work page 2025

  22. [22]

    Accessed: 2025-08-28

    URL https://www.ohchr.org/en/documents/country-reports /ahrc5923-economy-occupation-economy-genocide-report-special-rap porteur. Accessed: 2025-08-28. PBS NewsHour. Amid pressure from employees, Google drops Pentagon’s Project Maven account, June

    work page 2025

  23. [23]

    2026-02-

    URL https://www.pbs.org/newshour/show/amid-pressure-fro m-employees-google-drops-pentagons-project-maven-account . 2026-02-

    work page 2026

  24. [24]

    Accessed: 2025-03-07

    URL https://time.com/7013685/google-ai-deepmind-military-con tracts-israel. Accessed: 2025-03-07. RCW. Reaching critical will,

    work page arXiv 2025

  25. [25]

    Accessed: 2026-02-05

    URL https://reachingcriticalwill.org. Accessed: 2026-02-05. Michael Rothberg.The implicated subject: Beyond victims and perpetrators. Stanford University Press,

    work page 2026

  26. [26]

    Accessed: 2026-02-05

    URL https://scienceforthepeople.org. Accessed: 2026-02-05. Riley Simmons-Edler, Ryan Badman, Shayne Longpre, and Kanaka Rajan. AI-powered autonomous weapons risk geopolitical instability and threaten AI research.arXiv preprint arXiv:2405.01859,

    work page arXiv 2026

  27. [27]

    Accessed: 2026-02-05

    URL https://www.stopkillerrobots.org . Accessed: 2026-02-05. Haico Te Kulve and Wim A Smit. Civilian–military co-operation strategies in developing new technologies.Research policy, 32(6):955–970,

    work page 2026

  28. [28]

    Accessed: 2025-03-07

    URL https://www.washingtonpost.com/technology/2024/12/04/openai-and uril-military-ai. Accessed: 2025-03-07. Joseph Weizenbaum. Computers in uniform. A good fit?Science for the People, vol. 17, no.1&2, Computing the Future,

    work page 2024

  29. [29]

    Basic research, lethal effects: Military AI research funding as enlistment.arXiv preprint arXiv:2411.17840,

    David Gray Widder, Sireesh Gururaja, and Lucy Suchman. Basic research, lethal effects: Military AI research funding as enlistment.arXiv preprint arXiv:2411.17840,

    work page arXiv

  30. [30]

    Accessed: 2025-03-07

    URL https://techcrunch.com/2024/11/07/anthropic-teams-up-with-palan tir-and-aws-to-sell-its-ai-to-defense-customers. Accessed: 2025-03-07. 9

    work page 2024