Project Maven Research Articles

AI-Centric secure outer space operations

This article critically assesses the transformative role of Artificial Intelligence (AI) in military operations, focusing on terrestrial warfare and outer space security. It offers four main points of discussion: 1) An evaluation of AI applications in terrestrial warfare, using real-world technologies such as Project Maven and BAE Systems’ Taranis; 2) an examination of AI’s contributions and risks in the field of cybersecurity; 3) an in-depth look at AI’s growing influence in space security, including technical aspects of systems like the U.S. Space-Based Infrared and the European Data Relay System; 4) an analysis of the ethical and policy challenges associated with AI deployment, informed by the author’s viewpoints on the necessity of international regulation. Drawing from various case studies and expert consultations, the article highlights AI’s capabilities in enhancing decision-making and operational efficiency while discussing the ethical and technical complexities it introduces. The study concludes by offering nuanced recommendations for responsibly integrating AI into military strategies and policies, especially those concerning space security. The aim is to inform and guide military professionals and policymakers by providing actionable insights for responsible decision-making in AI-augmented conflict scenarios on Earth and in space.

Ethics as an enabler and a constraint – Narratives on technology development and artificial intelligence in military affairs through the case of Project Maven

Project Maven is an AI-induced information technology for military applications initiated by the United States Department of Defence (DoD) in 2017 and originally signed on to a civilian contractor, namely Google. However, this initiative raised massive resistance from a substantial amount of Google employees, eventually leading to the contract's annulation. This article uses narrative analysis to investigate enabling and constraining arguments of AI for military purposes that appeared in the debate following the public announcement of Project Maven. In addition, the article highlights the co-production of ethics, technology, and the complex issues that arise from civilian-military exchanges in technology development. Enabling arguments associated with consequentialist ethics are identified as narratives of accuracy and maintenance. Accuracy constitutes a guiding principle for saving civilian lives, while maintenance is directed at keeping the power balance intact. In contrast, constraining arguments proceed from deontological ethics that emphasize disengagement and ambivalence. Disengagement amplifies a civilian/military divide, while ambivalence exhibits conflicting views concerning the prospect of supplementing technological solutions that have the potential to contribute to war and civilian casualties. Conclusively, security narratives and technological storytelling are important aspects to consider since they hold a performative function that influences the framing and mobilization of security and technology development.

Evaluation of Simulated Recognition Aids for Human Sensemaking in Applied Surveillance Scenarios.

We determined whether the human capability for sensemaking, or for identifying essential elements of information (EEIs), could be enhanced by a simulated recognition aid that directed attention to people and vehicles in scenes or by a simulated recognition aid that directed attention to EEIs. For intelligence analysts, sensemaking is challenging because it frequently involves making inferences about uncertain data. One way to enhance sensemaking may involve collaboration from a machine recognition aid such as Project Maven, an established algorithm that directs analysts' attention to people and vehicles in scenes. We simulated the directed attention of Project Maven as well as a machine recognition aid that directed attention to EEIs. We created full-motion videos of simulated compounds viewed by an overhead camera. Sensemaking was assessed by measuring participants' ability to predict events and identify signs. Participants' attention was directed by placing small globe symbols above either all people and vehicles, or all EEIs. Novices and intelligence analysts participated. Simulated recognition aids directing participants' attention to EEIs improved EEI identification but directing attention to people and vehicles (emulating Project Maven) did not. Participants' sensemaking was not enhanced by either type of simulated recognition aid. Guiding attention to features in a scene improves their identification whereas indiscriminate steering of attention to entities in the scene does not improve understanding of the holistic meaning of events, unless attention is drawn to relevant signs of those events. Results contribute to our goal of determining which human-machine systems improve the sensemaking capability of intelligence analysts in the field.

Machine Learning and the Platformization of the Military: A Study of Google's Machine Learning Platform TensorFlow

Abstract Against the background of the growing use of machine learning (ML) based technologies by the military, our article calls for an analytical perspective on ML platforms to understand how ML proliferates across the military and to what effects. Adopting a material–technical perspective on platforms as developed within new media studies, and bringing this literature to critical security studies, we suggest that a focus on platforms and the technical work they do is needed to understand how digital technologies are emerging and shaping security practices. Through a detailed study of Google's open-source ML platform TensorFlow and a discussion of the US Department of Defense Algorithmic Warfare Cross-Functional Team, or Project Maven, we make two broader contributions. First, we identify a broader “platformization” of the military, with which we refer to the growing involvement and permeation of the (technomaterial) ML platform as the infrastructure that enables new practices of decentralized and experimental algorithm development across the military. Second, we draw out how this platformization is accompanied by new entanglements between the military and actors in the corporate domain, especially Big Tech, which play a key role in this context, as well as the open-source community that is organized around these platforms.

Ethics as an Enabler and a Constraint – Narratives on Technology Development and Artificial Intelligence in Military Affairs Through the Case of Project Maven

Ethics as an Enabler and a Constraint – Narratives on Technology Development and Artificial Intelligence in Military Affairs Through the Case of Project Maven

Corporate Governance of Artificial Intelligence in the Public Interest

Corporations play a major role in artificial intelligence (AI) research, development, and deployment, with profound consequences for society. This paper surveys opportunities to improve how corporations govern their AI activities so as to better advance the public interest. The paper focuses on the roles of and opportunities for a wide range of actors inside the corporation—managers, workers, and investors—and outside the corporation—corporate partners and competitors, industry consortia, nonprofit organizations, the public, the media, and governments. Whereas prior work on multistakeholder AI governance has proposed dedicated institutions to bring together diverse actors and stakeholders, this paper explores the opportunities they have even in the absence of dedicated multistakeholder institutions. The paper illustrates these opportunities with many cases, including the participation of Google in the U.S. Department of Defense Project Maven; the publication of potentially harmful AI research by OpenAI, with input from the Partnership on AI; and the sale of facial recognition technology to law enforcement by corporations including Amazon, IBM, and Microsoft. These and other cases demonstrate the wide range of mechanisms to advance AI corporate governance in the public interest, especially when diverse actors work together.

Analyzing Dilemmas Posed by Artificial Intelligence and 4IR Technologies Requires using all Available Models, Including the Existing International Human Rights Framework and Principles of AI Ethics

We are living in the epoch referred to as the ‘4th industrial revolution'. The 4th industrial Revolution (4IR) is a development characterized by a fusion of technologies that blur the digital, physical, and biological spheres (e.g., cyberspace, virtual and augmented reality, body-machine interface and robotics). Certain is the guaranteed ubiquitous adoption of these technologies, and futurism. Where the former is a reference to the increasing use and normalization of such technologies in everyday life, government service provision and industry. The latter is a reference to the philosophical/science fiction discussions that are emerging as a result of these changes (e.g. debates around the ‘singularity’, transhumanism, and posthumanism – often presented in utopian/dystopia terms). As such, the definition of digital ethics can be expanded and expressed in terms of the impacts of new digital technologies, through analysis of potential opportunities and risks in contemporary and future contexts. Many are working on forward‑looking policy frameworks and governance protocols, with broad multistakeholder engagement and buy‑in, to accelerate the adoption of emerging technologies in the global public interest, such as artificial intelligence (AI) and machine learning (ML) blockchain, 5G, data analytics, quantum computing, autonomous vehicles, synthetic biology, the internet of things (IoT), and killer robots or autonomous weapons systems (AWS). We have gained insight into the unequal distribution of the positive and negative impacts of AI on human rights throughout society, and have begun to explore the power of the human rights framework to address these disparate impacts. Although internationally recognized laws and standards on human rights provide a common standard of achievement for all people in all countries, more work is needed to understand how they can be best applied in the context of disruptive technology. AI systems raise myriad questions for society and democracy, only some of which are covered or addressed by existing laws. In order to fill these perceived gaps, a vocal group of governments, industry players, academics, and civil society actors have been promoting principles or frameworks for ethical AI. COVID-19 accelerated the use of AI in all countries and all fields. The pandemic accelerated the transition to a society that is increasingly based on the use of AI. This also increased the threats new risks related to human rights in the context of AI deployment. The human rights implications of governments' aggressive measures targeting the spread of COVID-19-related misinforation is also discussed. The question of whether corporations can act ethically is particularly relevant for Big Tech. Many of these firms are oligopolies that individuals and governments alike depend on completely, though they have little to no capacity to independently remedy issues when they arise, as Project Maven showed. Artificial intelligence and automated decision-making tools are increasing in power and centrality, and technology companies retain large troves of private data that it sells. These companies are at the forefront of technological innovation and may be caught up with the factual question of what can be done rather than the normative question of whether it should be done. All these issues arise in a field where there is little to no government regulation or intervention. The threats AI poses to society are so new, that the legal system is struggling to impose sufficient values and restrictions. Thus, a coherent approach to addressing AI ethics, values and consequences is, indeed, urgently needed. In May 2019, 42 countries adopted the Organization for Economic Co-operation and Development (OECD) AI Principles, a legal recommendation that includes five principles and five recommendations related to the use of AI. To ensure the successful implementation of the Principles, the OECD launched the AI Policy Observatory in February 2020. The Observatory publishes practical guidance about how to implement the AI Principles, and supports a live database of AI policies and initiatives globally. It also compiles metrics and measurement of global AI development and uses its convening power to bring together the private sector, governments, academia, and civil society. The AI ethics and governance initiatives discussed are cause for optimism that the global community will use all available models and brainpower for analysis and ultimately global governance of AI.

Algorithmic warfare and the reinvention of accuracy

ABSTRACT This article aims to integrate two interrelated strands in critical security studies. The first is mounting evidence for the fallacy of claims for precision and accuracy in the United States ‘counterterrorism’ programme, particularly as it involves expanding aerial surveillance in support of operations of extrajudicial assasination. The second line of critical analysis concerns growing investment in the further automation of these operations, more specifically in the form of the US Department of Defense Algorithmic Warfare Cross-Functional Team, or Project Maven. Building upon generative intersections of critical security studies and science and technology studies (STS), I argue that the promotion of automated data analysis under the sign of artificial intelligence can only serve to exacerbate military operations that are at once discriminatory and indiscriminate in their targeting, while remaining politically and legally unaccountable.

Negotiating 'Evil': Google, Project Maven and the Corporate Form

‘Don’t be evil’ was part of Google’s corporate code of conduct since 2000; however, it was quietly removed in April or May 2018 and subsequently replaced with ‘do the right thing’. Questions were raised both internally and externally to the organisation regarding the substantive meaning of this imperative. Some have highlighted the company’s original intentions in creating the code of conduct, while others have used the motto as a basis for critiquing the company—such as for its advertising practices, failure to pay corporate tax or the manipulation of Google-owned content. The imperative’s removal occurred at a time when thousands of Google employees, including senior engineers, signed a letter protesting the company’s involvement in Project Maven, a Pentagon program that uses artificial intelligence to interpret video imagery, which could in turn be used to improve the targeting capability of drone strikes. Employees asserted their refusal to be involved in the business of war and expressed their wariness of the United States government’s use of technology. This article will examine the legal construct and concept of the corporation, and whether it is possible for corporations to not be evil in the twenty-first century.

Epstein, Project Maven, and Some Reasons to Think About Where We Get Our Funding [Ethical, Legal, and Societal Issues

Epstein, Project Maven, and Some Reasons to Think About Where We Get Our Funding [Ethical, Legal, and Societal Issues

AI, Machine Learning & Deep Learning Risk Management & Controls: Beyond Deep Learning and Generative Adversarial Networks: Model Risk Management in AI, Machine Learning & Deep Learning

The current paper proposes how model risk management in operationalizing machine learning for algorithm deployment can be applied in national C4I and Cyber projects such as Project Maven. It builds upon recent leadership of global Management and Leadership industry executives for AI and Machine Learning Executive Education for MIT Sloan School of Management and the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL) and invited presentations at Princeton University. After building understanding about why model risk management is most crucial to robust AI, Machine Learning, Deep Learning, and, Neural Networks deployment, it introduces a Knowledge Management Framework for Model Risk Management to advance beyond ‘AI Automation’ to ‘AI Augmentation.’

Cognitive Computing for Anticipatory Risk Analytics in Intelligence, Surveillance, & Reconnaissance (ISR): Model Risk Management in Artificial Intelligence & Machine Learning (Presentation Slides)

Drawing upon insights shared in the MIT: AI & Machine Learning: Management and Leadership learning community of practice, the current Intelligence, Surveillance, Reconnaissance (ISR) presentation advances the focus on collective intelligence of people and computers in the context of Cognitive Computing for Anticipatory Risk Analytics in Intelligence, Surveillance, & Reconnaissance (ISR). It defines as well as distinguishes multi-level Cognitive Computing process engineering frameworks of Artificial Intelligence (AI) & Machine Learning as applied in KM practices of US and worldwide firms, governments, and ISR agencies from Cognitive-Neuromorphic Chips. A recent IEEE Spectrum report also published as Chip's Make-or-Break Moment observes that Neuromorphic Chips Are Destined for Deep Learning—or Obscurity given that the Chip researchers have hitched their wagon to deep learning's star. Drawing upon insights on Model Risk Management and Anticipatory Risk Analytics focus of top Wall Street investment banks and hedge funds beyond the Global Financial Crisis currently guiding national and global Cyber Risk Insurance industry practices, we demonstrate how Model Risk Management (MRM) and Anticipatory Risk Analytics are even more critical in the global and national domains of Intelligence, Surveillance, Reconnaissance (ISR). The first operational use of AI and Deep Learning AI technologies in the Defense Intelligence Enterprise led by Project Maven and Algorithmic Warfare Cross-Functional Team is used as a case study for illustrating how Anticipatory Risk Analytics & MRM assume even greater significance at the intersections of Space and Cyberspace wherein Offensive and Defensive Cybersecurity strategies, such as discussed in the recent 2015 and 2016 presentations at the Princeton Quant Trading Conference, are deployed.