Agent-Supported Foresight for AI Systemic Risks: AI Agents for Breadth, Experts for Judgment

AI impact assessments often stress near-term risks because human judgment degrades over longer horizons, exemplifying the Collingridge dilemma: foresight is most needed when knowledge is scarcest. To address long-term systemic risks, we introduce a scalable approach that simulates in-silico agents using the strategic foresight method of the Futures Wheel. We applied it to four AI uses spanning Technology Readiness Levels (TRLs): Chatbot Companion (TRL 9, mature), AI Toy (TRL 7, medium), Griefbot (TRL 5, low), and Death App (TRL 2, conceptual). Across 30 agent runs per use, agents produced 86-110 consequences, condensed into 27-47 unique risks. To benchmark the agent outputs against human perspectives, we collected evaluations from 290 domain experts and 7 leaders, and conducted Futures Wheel sessions with 42 experts and 42 laypeople. Agents generated many systemic consequences across runs. Compared with these outputs, experts identified fewer risks, typically less systemic but judged more likely, whereas laypeople surfaced more emotionally salient concerns that were generally less systemic. We propose a hybrid foresight workflow, wherein agents broaden systemic coverage, and humans provide contextual grounding. Our dataset is available at: https://social-dynamics.net/ai-risks/foresight.

• Publication year

  2026

• Venue

  Unknown venue

• Publication date

  2026-02-09

• Fields of study

  Computer Science

• Identifiers
  arXiv 2602.08565
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Seeking Inspiration through Human-LLM Interaction

  2025cited by this paper
• Initiating the Global AI Dialogues: Laypeople Perspectives on the Future Role of genAI in Society from Nigeria, Germany and Japan

  2025cited by this paper
• The impacts of companion AI on human relationships: risks, benefits, and design considerations

  2025cited by this paper
• RiskRAG: A Data-Driven Solution for Improved AI Model Risk Reporting

  2025cited by this paper
• AI Risk Atlas: Taxonomy and Tooling for Navigating AI Risks and Resources

  2025cited by this paper
• The Time Machine: Future Scenario Generation Through Generative AI Tools

  2025cited by this paper
• Understanding the Impacts of Generative AI Use on Children: WP2 School based Engagements

  2025cited by this paper
• Digital Dybbuks and virtual Golems: the ethics of digital duplicates in Holocaust testimony

  2025cited by this paper
• CoExploreDS: Framing and Advancing Collaborative Design Space Exploration Between Human and AI

  2025cited by this paper
• AI Risk Assessment: A Scenario-Based, Proportional Methodology for the AI Act

  2024cited by this paper
• Exploring the Integration of Artificial Intelligence in Delphi Studies: A Comparative Analysis of Human and AI Expert Panels

  2024cited by this paper
• From Prediction to Foresight: The Role of AI in Designing Responsible Futures

  2024cited by this paper
• A Taxonomy of Systemic Risks from General-Purpose AI

  2024influential reference
• "Just Like, Risking Your Life Here": Participatory Design of User Interactions with Risk Detection AI to Prevent Online-to-Offline Harm Through Dating Apps

  2024cited by this paper
• The Dark Side of AI Companionship: A Taxonomy of Harmful Algorithmic Behaviors in Human-AI Relationships

  2024cited by this paper
• Plurals: A System for Guiding LLMs via Simulated Social Ensembles

  2024influential reference
• PersonaFlow: Designing LLM-Simulated Expert Perspectives for Enhanced Research Ideation

  2024cited by this paper
• Co-designing an AI Impact Assessment Report Template with AI Practitioners and AI Compliance Experts

  2024cited by this paper
• Supermind Ideator: How scaffolding Human-AI collaboration can increase creativity

  2024cited by this paper
• ExploreGen: Large Language Models for Envisioning the Uses and Risks of AI Technologies

  2024cited by this paper
• The Good, The Bad, and The Greedy: Evaluation of LLMs Should Not Ignore Non-Determinism

  2024cited by this paper
• Good Intentions, Risky Inventions: A Method for Assessing the Risks and Benefits of AI in Mobile and Wearable Uses

  2024cited by this paper
• BLIP: Facilitating the Exploration of Undesirable Consequences of Digital Technologies

  2024cited by this paper
• Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems

  2024cited by this paper
• Risk and Response in Large Language Models: Evaluating Key Threat Categories

  2024cited by this paper
• Farsight: Fostering Responsible AI Awareness During AI Application Prototyping

  2024cited by this paper
• A Collaborative, Human-Centred Taxonomy of AI, Algorithmic, and Automation Harms

  2024cited by this paper
• AI Risk Categorization Decoded (AIR 2024): From Government Regulations to Corporate Policies

  2024cited by this paper
• CharacterMeet: Supporting Creative Writers' Entire Story Character Construction Processes Through Conversation with LLM-Powered Chatbot Avatars

  2024cited by this paper
• A global scale comparison of risk aggregation in AI assessment frameworks

  2024cited by this paper
• Generative AI enhances individual creativity but reduces the collective diversity of novel content

  2023cited by this paper
• An Overview of Catastrophic AI Risks

  2023cited by this paper
• Taking AI risks seriously: a new assessment model for the AI Act

  2023cited by this paper
• The Participatory Turn in AI Design: Theoretical Foundations and the Current State of Practice

  2023cited by this paper
• Booklet-Based Design Fiction to Support AI Literacy

  2022cited by this paper
• Corporate foresight: A systematic literature review and future research trajectories

  2022cited by this paper
• Sources of Risk of AI Systems

  2022cited by this paper
• Toward a meta-theory of creativity forms: How novelty and usefulness shape creativity

  2022influential reference
• Participatory Design of AI Systems: Opportunities and Challenges Across Diverse Users, Relationships, and Application Domains

  2022cited by this paper
• Actionable Guidance for High-Consequence AI Risk Management: Towards Standards Addressing AI Catastrophic Risks

  2022cited by this paper
• How to visualise futures studies concepts: Revision of the futures cone

  2022cited by this paper
• Social Perception of Systemic Risks

  2021cited by this paper
• Innovative idea generation in problem finding: Abductive reasoning, cognitive impediments and the promise of Artificial Intelligence

  2021cited by this paper
• Anticipating and addressing the ethical implications of deepfakes in the context of elections

  2020influential reference
• Judgment Call the Game: Using Value Sensitive Design and Design Fiction to Surface Ethical Concerns Related to Technology

  2019cited by this paper
• Forensic architecture

  2019cited by this paper
• Reliability and Inter-rater Reliability in Qualitative Research

  2019cited by this paper
• Figma: collaborative interface design tool

  2019influential reference
• Intimate Futures: Staying with the Trouble of Digital Personal Assistants through Design Fiction

  2018cited by this paper
• Eliciting Values Reflections by Engaging Privacy Futures Using Design Workbooks

  2017cited by this paper
• From the Perspective of Capability: Identifying Six Roles for a Successful Strategic Foresight Process

  2016cited by this paper
• Grounded theory

  2015cited by this paper
• Normal Accidents Living with High-Risk Technologies

  2012cited by this paper
• Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory

  2010cited by this paper
• A methodology for evaluating technology readiness during product development

  2009cited by this paper
• Envisioning systemic effects on persons and society throughout interactive system design

  2008cited by this paper
• Thinking about the Future: Guidelines for Strategic Foresight

  2008cited by this paper
• Construal level theory

  2007cited by this paper
• Value scenarios: a technique for envisioning systemic effects of new technologies

  2007cited by this paper
• “If I look at the mass I will never act”: Psychic numbing and genocide

  2007cited by this paper
• Identifying Quality, Novel, and Creative Ideas: Constructs and Scales for Idea Evaluation

  2006influential reference
• A Mathematical Theory of Communication

  2006cited by this paper
• Futures research methodology—Version 2.0

  2004influential reference
• Golden Eggs and Hyperbolic Discounting

  1997cited by this paper
• The Logic Of Failure: Recognizing And Avoiding Error In Complex Situations

  1997cited by this paper
• Beyond Accuracy: What Data Quality Means to Data Consumers

  1996cited by this paper
• Qualitative Data Analysis: A Methods Sourcebook

  1994cited by this paper
• Exploring the "planning fallacy": Why people underestimate their task completion times.

  1994cited by this paper
• Advances in prospect theory: Cumulative representation of uncertainty

  1992cited by this paper
• Judgment Under Uncertainty: Heuristics and Biases.

  1984cited by this paper
• Specious reward: a behavioral theory of impulsiveness and impulse control.

  1975cited by this paper
• Futurizing Teaching vs. Futures Courses.

  1972cited by this paper
• Counterintuitive behavior of social systems

  1971cited by this paper
• You have printed the following article : Prospect Theory : An Analysis of Decision under Risk

  year unknowncited by this paper

• No citing papers are available for this paper.