Artificial Intelligence and New Threat Vectors: Using Scenario Planning for Trend Forecasting

Ann Fitzgerald, Dmytro Chumachenko, Halyna Padalko, Vijay GaneshBalsillie 2023 · DOI

Abstract

As AI technology evolves rapidly, concerns over its potential threats and the necessity for regulatory mechanisms have become central to the global conversation. This paper proposes a new framework for classifying AI-driven disinformation threats into three categories: fake identities, fake information, and fake behavior. By applying "Field Anomaly Relaxation" (FAR) scenario planning, the authors demonstrate how these vectors intersect to create "wicked problems" that threaten democratic stability, national security, and economic opportunity.

Introduction: The Existential Debate

The rapid advancement of artificial intelligence has moved beyond technical circles into the heart of public and political discourse. The debate often centers on whether AI poses an existential threat to humanity, reflecting a deep-seated anxiety about the technology's autonomous capabilities and the speed of its integration into society.

Public debate on the existential threat of AI research and development.

This paper addresses the "institutional lag"—the widening gap between the speed of AI innovation and the ability of governance structures to regulate it. The authors argue that AI is no longer just a tool but an entity capable of sophisticated "information warfare," necessitating a shift in how we perceive and prepare for emerging threats.

The Triad Model of Disinformation

To move beyond siloed policy responses, the authors introduce a systematic classification for AI-driven threats. This model identifies three primary vectors that interact to amplify disinformation campaigns:

  1. Fake Identities: The use of deepfakes and synthetic personas to create untraceable or deceptive influencers.
  2. Fake Information: The generation of biased, discriminatory, or entirely fabricated content via Large Language Models (LLMs).
  3. Fake Behaviour: The deployment of automated bot networks and social engineering to manipulate online discourse and create artificial echo chambers.
The interconnected cycle of fake identities, fake information, and fake behaviour.

The intersection of these vectors creates "wicked problems"—complex, interconnected challenges that defy simple, single-domain solutions. For example, a single campaign might combine a fake opinion leader (Identity) with biased, AI-generated content (Information) amplified by a bot-driven echo chamber (Behavior).

Methodology: Field Anomaly Relaxation (FAR)

The paper utilizes Field Anomaly Relaxation (FAR), a variant of morphological analysis, to forecast potential futures. Unlike traditional forecasting that looks at linear "best" or "worst" cases, FAR maps out "constellations" of plausible scenarios by linking factors across the three threat vectors. This allows researchers and policymakers to visualize the non-linear ways these technologies can be weaponized.

Strategic Insights and Scenarios

The FAR analysis yielded four primary scenarios that highlight the multifaceted nature of AI threats:

  • Scenario 1: Discrediting influential leaders through untraceable deepfakes and disinformation.
  • Scenario 2: Intensifying political polarization by manipulating biased and discriminatory content.
  • Scenario 3: Shifting public dialogue via fake opinion leaders supported by inauthentic follower networks.
  • Scenario 4: Financial fraud and data breaches driven by AI-powered identity theft and social engineering.

Beyond these scenarios, the authors identify critical systemic risks, including a "leadership deficit." As AI-driven character assassination becomes cheaper and more effective, the personal cost of entering public leadership escalates, potentially leading to a talent drain in democratic institutions. Furthermore, the paper suggests that data should no longer be viewed just as a commodity, but as a "national instrument of power," requiring its own dedicated government institutions.

Conclusion: Addressing the Knowledge Deficit

The greatest vulnerability facing modern democracies is the "knowledge and skill set deficit" at the intersection of technology and policy. Current academic and merit systems often reward narrow, single-domain expertise, which is insufficient for managing the multidisciplinary challenges of AI.

The authors call for transformational reforms in education and the development of international coalitions to share best practices. Only by closing the institutional lag and fostering a workforce capable of cross-domain analysis can governments hope to secure democratic stability and the social contract in the age of AI.

Summary updated

Abstract

Ann Fitz-Gerald, Dmytro Chumachenko, Halyna Padalko and Vijay GaneshAs AI technology evolves rapidly, concerns over its potential threats and the necessity for regulatory mechanisms have become central to the global conversation.
BibTeX
@inproceedings{fitzgerald2023artificial,
  author = {Ann Fitzgerald and Dmytro Chumachenko and Halyna Padalko and Vijay Ganesh},
  title = {{Artificial Intelligence and New Threat Vectors: Using Scenario Planning for Trend Forecasting}},
  booktitle = {Balsillie},
  year = {2023},
  doi = {10.51644/bap56},
}
Figure