Improving human collective decision-making through animal and artificial intelligence

The variability in responses generated by Large Language Models (LLMs) like OpenAI’s GPT-4 poses challenges in ensuring consistent accuracy on medical knowledge assessments, such as the United States Medical Licensing Exam (USMLE). This study introduces a novel multi-agent framework—referred to as a "Council of AIs"—to enhance LLM performance through collaborative decision-making. The Council consists of multiple GPT-4 instances that iteratively discuss and reach consensus on answers facilitated by a designated "Facilitator AI." This methodology was applied to 325 USMLE questions across Step 1, Step 2 Clinical Knowledge (CK), and Step 3 exams. The Council achieved consensus responses that were correct 97%, 93%, and 94% of the time for Step 1, Step 2CK, and Step 3, respectively, outperforming single-instance GPT-4 models. In cases where there wasn’t an initial unanimous response, the Council of AI deliberations achieved a consensus that was the correct answer 83% of the time. For questions that required deliberation, the Council corrected over half (53%) of responses that majority vote had gotten incorrect. At the end of deliberation, the Council often corrected majority responses that were initially incorrect: the odds of a majority voting response changing from incorrect to correct were 5 (95% CI: 1.1, 22.8) times higher than the odds of changing from correct to incorrect after discussion. We additionally characterized the semantic entropy of the response space for each question and found that deliberations impact entropy of the response space and steadily decrease it, consistently reaching an entropy of zero in all instances. This study showed that in a Council model response variability—often viewed as a limitation—could be leveraged as a strength, enabling adaptive reasoning and collaborative refinement of answers. These findings suggest new paradigms for AI implementation and reveal diversity of responses as a strength in collective decision-making even in medical question scenarios where there is a single correct response.Author SummaryIn our study, we explored how collaboration among multiple artificial intelligence (AI) systems could improve accuracy on medical licensing exams. While individual AI models like GPT-4 often produce varying answers to the same question—a challenge known as "response variability"—we designed a "Council of AIs" to turn this variability into a strength. The Council consists of several AI models working together, discussing their answers through an iterative process until they reach consensus.When tested on 325 medical exam questions, the Council achieved 97%, 93%, and 94% accuracy on the Step 1, Step 2CK, and Step 3, respectively. This improvement was most notable when answers required debate: in cases where initial responses disagreed, the collaborative process corrected errors 83% of the time. Our findings suggest that collective decision-making— even among AIs—can enhance accuracy and AI collaboration can potentially lead to more trustworthy tools for healthcare, where accuracy is critical. By demonstrating that diverse AI perspectives can refine answers, we challenge the notion that consistency alone defines a "good" AI. Instead, embracing variability through teamwork might unlock new possibilities for AI in medicine and beyond. This approach could inspire future systems where AIs and humans collaborate (e.g. on Councils with both humans and AIs), combining strengths to solve complex problems. While technical challenges remain, our work highlights a promising path toward more robust, adaptable AI solutions.

Group decision-making, in which a collective decision needs to be derived from individual agents’ preferences has been an active area of research. In group decisionmaking, autonomous agents need some procedures or mechanisms that can collect preference information from them and make a decision; or some protocols that enable them to be involved in the decision process, to interact with each other and to jointly make a collective decision. Examples include preference aggregation mechanisms, negotiation protocols, voting procedures, and auctions. This thesis studies the problem of agent-based group decision-making in various problem settings, proposing several efficient approaches to support multiple agents in reaching efficient and fair collective decisions overmultiple issues. The research work in this thesis can be divided into two parts, addressing the agent-based group decision-making problem in circumstances in which there is, respectively, complete and incomplete information available. The first part of this thesis focuses on preference aggregation in combinatorial domains, provided all agents’ preferences. Structured preferences Conditional Preference Network (CP-nets) and its variant Trade-off enhanced CP-nets (TCP-nets) are used as formal models for representing the agents’ preferences in group decision-making. Before going into the problem of making a group decision from a collection of CP-nets or TCP-nets, the problem of individual preference reasoning is studied and a computationally efficient heuristic algorithm for dominance testing in CP-nets, which forms the basis for many group decision-making mechanisms, is introduced. Subsequently, we consider the group decision-making problem with multiple agents’ CP-nets and TCP-nets. As the space of possible outcomes in combinatorial domains has a size exponential in the number of variables, computational complexity in group decision-making will be one of the issues that need to be addressed. Two computationally efficient approaches are proposed in order to support agent-based group decisionmaking in combinatorial domains with different decision criteria: (i) an approach that converts the qualitative preferences derived from CP-nets or TCP-nets into numerical penalty scoring functions, and then searches for an optimal collective decision based on the individual penalty scores; and (ii) an approach that focuses on combinatorial vote and computes the winners (i.e., winning alternatives) from a collection of CP-nets according to Majority voting rule. The second part of this thesis is concerned with the disadvantage of disclosing preference information in group decision-making and considers the incomplete information setting. Two negotiation approaches are presented in order to support agent-based group decision-making in, respectively, combinatorial domains and in domains with multiple continuous issues. Firstly, a negotiation protocol in combinatorial domains is proposed, which enables agents to distributively make decisions and leads them to Pareto-optimal agreements under incomplete information setting. Then for the domains with multiple continuous issues, a cooperative mediated negotiation approach is proposed to support the agents searching for joint utility gains and reaching an efficient and fair agreement. For every approach introduced in this thesis, experiments are carried out and the results are presented and analysed in order to provide insights into the performance of these approaches.

Artificial Intelligence (AI) is a transformative field at the intersection of computer science, mathematics, and cognitive psychology.It involves the development of systems capable of performing tasks that typically require human intelligence, such as learning, problem-solving, perception, and decision-making.AI technologies, including machine learning, natural language processing, and robotics, are revolutionizing industries ranging from healthcare to finance, and reshaping the way we live, work, and interact with technology."In the 21st century, Artificial Intelligence (AI) has emerged as a transformative force reshaping various aspects of society and industry.With advancements in computing power, data availability, and algorithmic techniques, AI has made significant strides in enabling machines to perform tasks traditionally requiring human intelligence.One of the key drivers of AI progress in the 21st century has been machine learning, a subfield of AI that focuses on algorithms capable of learning from data and making predictions or decisions.Deep learning, a subset of machine learning that uses artificial neural networks with many layers, has particularly fuelled breakthroughs in areas such as image and speech recognition, natural language processing, and autonomous vehicles.AI applications have become pervasive across industries, revolutionizing sectors such as healthcare, finance, transportation, manufacturing, and entertainment.In healthcare, AI is being used for disease diagnosis, drug discovery, personalized medicine, and medical image analysis.In finance, it's employed for fraud detection, algorithmic trading, and customer service.Autonomous vehicles and smart transportation systems leverage AI for navigation, route optimization, and traffic management.Moreover, AI-driven virtual assistants like Siri, Alexa, and Google Assistant have become ubiquitous, enhancing user experience and productivity.AIpowered recommendation systems personalize content delivery on platforms like Netflix, Spotify, and Amazon, improving user engagement and satisfaction.However, alongside its transformative potential, AI raises ethical, social, and economic concerns.Issues such as job displacement due to automation, algorithmic biases, data privacy, and the potential misuse of AI for surveillance or military purposes require careful consideration and regulation.Overall, the 21st century has witnessed AI evolve from a theoretical concept to a practical technology with profound implications for society, promising both opportunities and challenges as it continues to advance. Artificial Intelligence and Human Society:Artificial Intelligence (AI) is deeply intertwined with human society, impacting various aspects of our daily lives, work, and culture.Here's how AI influences human society:1. Workforce Automation: AI technologies automate routine and repetitive tasks across industries, leading to increased efficiency and productivity.However, concerns about job displacement and the need for reskilling and upskilling arise as AI replaces certain human roles.2. Healthcare: AI applications in healthcare facilitate disease diagnosis, personalized treatment plans, drug discovery, and medical imaging analysis.AI-driven predictive analytics also help in identifying health risks and optimizing patient outcomes.3. Education: AI enhances education through personalized learning platforms, intelligent tutoring systems, and adaptive assessments tailored to individual student needs.It also enables the development of educational content and resources based on learner preferences and performance.4. Communication and Interaction: AI-driven virtual assistants, chatbots, and language translation services improve communication and accessibility.Natural language processing enables human-like interactions with AI systems, enhancing user experience and accessibility for people with disabilities.

Artificial Intelligence (AI) has emerged as a transformative force in human society, offering both opportunities and challenges across various sectors. This paper explores the dual impact of AI on human society, focusing on its potential to enhance efficiency, improve decision-making, and drive economic growth, while also addressing the ethical, social, and economic challenges it poses. AI, defined as the simulation of human intelligence by machines, has rapidly evolved, enabling real-time data analysis, automation of repetitive tasks, and advancements in fields such as healthcare, education, and customer service. However, its integration into society raises significant concerns, including ethical dilemmas, privacy issues, job displacement, and threats to human security. The study employs a qualitative research design, analyzing existing literature to identify the positive and negative impacts of AI. This study applies the Innovation Adoption Theory to understand how AI technologies are integrated into society and the factors influencing their acceptance. The findings indicate that while AI holds immense promise for revolutionizing multiple facets of human life, its benefits must be carefully balanced against inherent risks. Further findings reveal that AI has revolutionized industries by improving healthcare diagnostics, enabling personalized learning, and enhancing customer service through automation and data-driven insights. Despite these benefits, the paper highlights the ethical challenges posed by AI, such as algorithmic biases, accountability in autonomous systems, and the potential for misuse in surveillance and warfare. Additionally, the widespread adoption of AI has led to job losses in traditional sectors, raising concerns about unemployment and economic inequality. The paper further discusses the future prospects of AI, emphasizing its potential to drive the Fourth Industrial Revolution. While AI technologies continue to advance, particularly in machine learning and deep learning, their long-term societal impact remains uncertain. The paper concludes that while AI offers significant benefits, its integration into society must be carefully managed to mitigate risks and ensure equitable outcomes. Policymakers, businesses, and researchers must collaborate to address the ethical, social, and economic challenges posed by AI, ensuring that its development aligns with the broader goals of human welfare and societal progress.

A major goal in evolutionary biology is to elucidate common principles that drive human and other animal societies to adopt either a warlike or peaceful nature. One proposed explanation for the variation in aggression between human societies is the democratic peace hypothesis. According to this theory, autocracies are more warlike than democracies because autocratic leaders can pursue fights for private gain. However, autocratic and democratic decision-making processes are not unique to humans and are widely observed across a diverse range of non-human animal societies. We use evolutionary game theory to evaluate whether the logic of democratic peace may apply across taxa; specifically adapting the classic Hawk-Dove model to consider conflict decisions made by groups rather than individuals. We find support for the democratic peace hypothesis without mechanisms involving complex human institutions and discuss how these findings might be relevant to non-human animal societies. We suggest that the degree to which collective decisions are shared may explain variation in the intensity of intergroup conflict in nature.

In recent years, social innovation has become increasingly influential in both scholarship and policy. It is the conceptual foundation for communitybased trusts, think tanks, corporate management practices and government funding programs in every continent, leading to a wide range of projects and international networks which recognize past failures of conventional service delivery to tackle poverty and social exclusion, and seek to promote new ways of doing things, grounded in the social relations and experiences of those in need. It is the great inspiration for many social movements, associations, bottomup initiatives to claim improvements in their human conditions, their community life and their place in society. It has found a home in policy at the highest level, for example in the US Whitehouse’s office for Social Innovation and civic Participation, through the creation of the national Secretariat for Solidarity economy in brazil and in the european commission’s Innovation Policy programmes. It has become a lead term for corporate social responsibility, business ethics and the revisiting of the role of social enterprise and the social economy in socioeconomic development. The growing importance of the idea reflects wide and profound dissatisfaction with recent directions and outcomes of ‘innovation’ in technology, markets, policy and governance systems, and particularly a sense – to remain polite – that the benefits of such innovations have not been distributed as generally or as equitably as they should (see Jessop et al., chapter 8). This also holds for changes in sociopolitical regimes. Social innovation as a governance change with more bottomup participation, protection of the rights of ‘common’ citizens and collective decisionmaking systems has indeed increasingly become a mirror to reflect on the consequences of macroinstitutional changes such as the privatization of banks and social services, deregulation of markets at the expense of the satisfaction of collective needs, the heralding of elite consumerism as a value system, etc. In much policy and management discourse, social innovation refers broadly to innovation in meeting social needs of, or delivering social benefits to, communities – the creation of new products, services, organizational structures or activities that are ‘better’ or ‘more effective’ than traditional public sector, philanthropic or marketreliant approaches in responding to social exclusion. Particularly successful forms of social innovation in this sense, which to many people represent ‘iconic’ examples, include microfinance and popular education – gamechanging initiatives which have travelled well beyond their original geographical and social contexts to find permanent institutional homes in the public services of many countries (see for example ashta et al., chapter 6; dubeux, chapter 22; Fernandes et al., chapter 29). equally important, and part

Decision-making animals can use slow-but-accurate strategies, such as making multiple comparisons, or opt for simpler, faster strategies to find a ‘good enough’ option. Social animals make collective decisions about many group behaviours including foraging and migration. The key to the collective choice lies with individual behaviour. We present a case study of a collective decision-making process (house-hunting ants, Temnothorax albipennis), in which a previously proposed decision strategy involved both quality-dependent hesitancy and direct comparisons of nests by scouts. An alternative possible decision strategy is that scouting ants use a very simple quality-dependent threshold rule to decide whether to recruit nest-mates to a new site or search for alternatives. We use analytical and simulation modelling to demonstrate that this simple rule is sufficient to explain empirical patterns from three studies of collective decision-making in ants, and can account parsimoniously for apparent comparison by individuals and apparent hesitancy (recruitment latency) effects, when available nests differ strongly in quality. This highlights the need to carefully design experiments to detect individual comparison. We present empirical data strongly suggesting that best-of-n comparison is not used by individual ants, although individual sequential comparisons are not ruled out. However, by using a simple threshold rule, decision-making groups are able to effectively compare options, without relying on any form of direct comparison of alternatives by individuals. This parsimonious mechanism could promote collective rationality in group decision-making.

This study investigates the integration of Artificial Intelligence (AI) as a co-pilot in aviation. As aviation increasingly embraces automation, concerns about safety, pilot roles, and ethical implications grow. The research emphasizes the necessity for human-AI collaboration in which AI enhances pilot capabilities without replacing them. Using a mixed-methods approach that includes literature review, case analysis, and stakeholder input, the study provides a model for safely integrating AI into flight operations. Key findings support that AI should assist rather than replace pilots, maintaining human oversight in high-stakes environments. The thesis concludes with policy and design recommendations for AI integration in aviation. This comprehensive study explores the transformative potential of Artificial Intelligence (AI) in modern commercial aviation, specifically emphasizing the concept of AI as a co-pilot. With rising operational complexity, global pilot shortages, and increasing demand for enhanced safety and efficiency, AI presents itself as a promising technological innovation. It has the capacity to provide real-time flight data interpretation, system anomaly alerts, autonomous recommendations, and task automation that collectively enhance situational awareness and reduce pilot workload. However, the expansion of AI usage in aviation also introduces critical concerns around pilot-AI interaction, explainability of AI systems, cybersecurity risks, and ethical responsibility in cases of failure. The research argues for a balanced integration model where AI supports—rather than supplants—human pilots. The idea of a "co-pilot AI" is conceptualized to highlight systems that work in tandem with human decision-makers, offering real-time insights while respecting human authority. Human-in-the-loop (HITL) design frameworks, explainable AI (XAI) principles, and trust-calibration models were central to the theoretical foundation of this research. A mixed-methods approach was adopted to strengthen the analysis, combining literature reviews, international aviation policy reviews, expert interviews, and comparative incident analysis. The findings of the study reveal a general consensus among professionals that while AI significantly enhances operational responsiveness and data processing capabilities, human oversight remains non-negotiable for moral judgment and adaptive decision-making during complex flight conditions. The research further identifies that system transparency, pilot familiarity, and regulatory clarity are essential for cultivating pilot trust in AI tools. It also highlights training gaps in existing aviation programs, where pilots are inadequately prepared to interface effectively with AI technologies. From a managerial and policy standpoint, the study recommends a strategic shift in training curricula to include AI-system literacy, greater investment in explainable interfaces, cross-industry collaboration for best practices, and harmonization of international regulatory standards. These measures, if implemented, will not only prevent technological overdependence but also build a more resilient and adaptive aviation ecosystem.

The introduction of Artificial Intelligence (AI) in our contemporary society imposes historically unique challenges for humankind. The emerging autonomy of AI holds unique potentials of eternal life of robots, AI and algorithms alongside unprecedented economic superiority, data storage and computational advantages. However, the introduction of AI to society also raises ethical questions. What is the social impact of robots, algorithms, blockchain and AI entering the workforce and our daily lives on the economy and human society? Should AI become eternal or is there a virtue in switching off AI at a certain point? If so, we may have to define a ‘virtue of killing’ and a ‘right to destroy’ that may draw from legal but also philosophical sources to answer the question how to handle the abyss of killing with ethical grace and fair style. In light of robots already having gained citizenship and being attributed as quasi-human under Common Law jurisdiction, should AI and robots be granted full citizen rights – such as voting rights? Or should we simply reap the benefits of AI and consider to define a democracy with different classes having diversified access to public choice and voting – as practiced in the ancient Athenian city state, which became the cradle of Western civilization and democratic traditions spread around the globe. Or should we legally justify AI slaves to economically reap their benefits, as was common in ancient Rome, which became the Roman Law legal foundation for Continental and some of Scandinavian Law traditions and which inspired very many different codifications around the world. Finally, we may also draw from the Code Napoleon, the French Code Civil established under Napoleon in 1804, which defined male and female into two classes of human with substantial right and power differences, and – to this day – accounts for one of the few documents that have influenced the whole world in legal and societal ways. In asking critical questions and unraveling the ethical boundary conditions of our future artificial world, the paper thereby takes a descriptive – afar from normative – theoretical angle targeted at aiding a successful introduction of AI into our contemporary workforce, democracy and society.

Risk preference during collective decision making: ant colonies make risk-indifferent collective choices

Extensive research on judicial politics has documented disparities in adjudication and biases in judging. Yet, lacking statistical methods to infer individual preferences from group decisions, existing studies have focused on courts publishing individual judges’ opinions, leaving a gap in understanding collegial courts that report only collective and unanimous (‘per curiam’) panel decisions. We introduce a statistical methodology to identify the most fitting decision-theoretic models for such collective decisions, infer judges’ individual preferences, and quantify the inconsistency in the courts’ decisions. This methodology is applicable in various small group decision-making contexts where group assignments are repeated and exogenous. Applying it to the Swiss appellate court for asylum appeals, where decisions are made in three-judge panels, we find that in 45 per cent of cases, the chair-as-dictator rule applies (rather than majority rule). Although judges’ preferences vary strongly with partisanship, the partially collective decision making of the panel moderates this heterogeneity.

Viability of decision-making systems in human and animal groups

Artificial Intelligence (AI) and Big Data Analytics (BDA) are increasingly recognized as transformative forces that are reshaping industries, governance, and human society in profound ways. AI, with its foundations in machine learning, deep learning, natural language processing, and predictive modelling, requires vast amounts of data to train algorithms and generate reliable insights. Complementing this, BDA provides the scale, diversity, and velocity of data processing required to uncover hidden patterns, correlations, and trends within complex datasets. Together, AI and BDA form a synergistic paradigm in which Big Data serves as the fuel and AI operates as the engine, driving intelligent, data-driven innovation across multiple domains. Globally, AI-BDA applications are enabling predictive diagnostics, telemedicine platforms, genomic-driven personalized treatments, and advanced medical imaging, thereby enhancing both the accessibility and quality of care. In India, where the healthcare system must serve more than 1.4 billion citizens with limited resources, these technologies can play a pivotal role in addressing systemic challenges such as inadequate rural access, shortages of medical professionals, real-time disease surveillance, and hospital resource optimization. International best practices, including the NHS AI Lab in the United Kingdom, Mayo Clinic initiatives in the United States, and Ping a Good Doctor in China, provide models that India can adapt and scale to its unique demographic and infrastructural context. Despite their vast potential, the adoption of AI and BDA faces several barriers. Data privacy and cybersecurity remain critical concerns, particularly with the emergence of new regulations and the sensitive nature of health-related information. Additionally, algorithmic bias and the uneven availability of digital infrastructure could exacerbate existing inequities if not adequately addressed. Looking forward, advancements in Edge AI, Explainable AI (XAI), and quantum-enhanced analytics are expected to extend the scope and reliability of AI-BDA systems. This paper argues that the integration of AI and BDA is not merely a technological advancement but a paradigm shifts in how societies innovate, make decisions, and deliver essential services. Specifically, it highlights the need for a human-centric, ethical, and affordable approach to healthcare innovation in India in order to unlock the full transformative potential of these technologies. Keywords: Artificial Intelligence, Big Data Analytics, Healthcare Innovation, Predictive Diagnostics, Ethical AI, Edge AI.

This paper critically analyzes the impact of artificial intelligence (AI) on Jean-Jacques Rousseau’s concept of the general will (GW), proposing a practical, three-dimensional approach to address the challenges posed by these technologies to democratic governance. The current dominance of automated systems controlled by a few digital giants (“Big Tech”) threatens to undermine collective and democratic decision-making. Building on Rousseau’s notion of the general will as a collective and sovereign expression of the common good, I identify four ways AI erodes the GW: (1) the privatization of the public debate; (2) the individualization of practices; (3) the erosion of personal and collective will; (4) the risk of technological despotism. I then discuss possible pathways to address these challenges and recover the sense of the GW. First, education and mediatization of AI issues would enable the (re)formation of the GW. Then, public participation platforms could allow its expression. Finally, coming back to Rousseau, I call for a renewal of the terms of the social contract through public participation in AI choices. By linking Rousseau’s democratic ideals with modern challenges, the paper presents actionable pathways to align AI technologies with democratic principles and collective representation.

The rapid development of digital technologies and the emergence of the digital economy have significantly transformed managerial decision-making processes. Digitalization affects all areas of enterprise activity, including data collection, processing, and analysis, which directly influences the speed, quality, and effectiveness of managerial decisions. Modern managers increasingly rely on digital tools, analytical platforms, decision support systems, artificial intelligence (AI), and large datasets, enabling more informed, timely, and flexible decision-making. These capabilities are essential for maintaining competitiveness in dynamic and uncertain business environments. This article analyzes the transformational features of managerial decision-making under digitalization, emphasizing theoretical foundations and practical applications. Effective managerial decisions are determined not only by their formulation but also by successful implementation, which ensures achievement of both strategic and operational objectives. Key contemporary aspects include digitalization and automation, real-time data-driven analysis, scenario modeling, AI integration, flexibility, and adaptability. These features enhance optimization, responsiveness, and overall managerial efficiency, particularly in enterprises such as bakeries, where production, financial, marketing, and personnel management must respond to fluctuating markets, regulatory changes, and consumer expectations.The classification of managerial decisions is considered with respect to management levels, problem complexity, personnel involvement, functional orientation, and decision-making conditions. Digital technologies facilitate process automation, data integration, collective decision-making, and transparency, enabling managers to make well-grounded, efficient, and adaptive choices. The study demonstrates that incorporating digital tools and AI in decision-making improves performance, strengthens competitiveness, and supports sustainable development of modern enterprises.