Taking AI risks seriously
A new assessment model for the AI Act
Bibliographic Data
| ID | 20397350 |
|---|---|
| Authors | Claudio Novelli (0000-0003-4193-8604, Zambon (Italy), corresponding author), Federico Casolari (0000-0003-3281-5426, Zambon (Italy)), Antonino Rotolo (0000-0001-5265-0660, Zambon (Italy)), Mariarosaria Taddeo (0000-0002-1181-649X, Turing Institute), Luciano Floridi (0000-0002-5444-2280, University of Oxford) |
| Year | 2024 |
| Volume | 39 |
| Issue | 5 |
| Pages | 2493-2497 |
| Publication date | 2024-10-01 |
| Open Access | Yes |
| Type | ARTICLE |
| Venue | AI & Society (JOURNAL) |
| Journal identifiers | ISSN: 0951-5666 • E-ISSN: 1435-5655 |
| Publisher | Springer Science and Business Media LLC (PUBLISHER) |
| DOI | 10.1007/s00146-023-01723-z |
| OpenAlex | W4384025135 |
| Language | EN |
| Citations received | 22 |
| References cited | 7 |
The EU Artificial Intelligence Act (AIA) defines four risk categories: unacceptable, high, limited, and minimal. However, as these categories statically depend on broad fields of application of AI, the risk magnitude may be wrongly estimated, and the AIA may not be enforced effectively. This problem is particularly challenging when it comes to regulating general-purpose AI (GPAI), which has versatile and often unpredictable applications. Recent amendments to the compromise text, though introducing context-specific assessments, remain insufficient. To address this, we propose applying the risk categories to specific AI scenarios, rather than solely to fields of application, using a risk assessment model that integrates the AIA with the risk approach arising from the Intergovernmental Panel on Climate Change (IPCC) and related literature. This integrated model enables the estimation of AI risk magnitude by considering the interaction between (a) risk determinants, (b) individual drivers of determinants, and (c) multiple risk types. We illustrate this model using large language models (LLMs) as an example
Business · Compromise · Computer security · Political science · Risk assessment · Risk management · Computer Science · Law · Risk Perception and Management
The European Union Artificial Intelligence Act
AI Peril as media-cultural formation
Chasing artificial intelligence in shared socioeconomic pathways
A Tough Balancing Act – The Evolving AI Governance in Korea
Two paths of balancing technology and ethics
Shared Minds
Addressing the potential social risk of self-reported data within a computational-intensive world
Social, legal, and ethical implications of AI-Generated deepfake pornography on digital platforms
Ethical Management of Ai-Produced Digital Works
AI product knowledge as moderator of trust and attitude in usage intention
AI’s Influence on Human Creative Agency
Destined for Balance? Centralized and Decentralized Approaches to AI Governance
Protecting democracy
Generative AI and digital resilience
Government regulation or industry self-regulation of AI? Investigating the relationships between uncertainty avoidance, people’s AI risk perceptions, and their regulatory preferences in Europe
Risk and artificial general intelligence
Leveraging teleological explanation to support general-purpose AI assessment
A Novel Type of Precautionary Argument for Situations of Severe Uncertainty in Science and Policy
On the Brussels-Washington Consensus About the Legal Definition of Artificial Intelligence
AI-Related Risk
We Need a Cern for AI”
AI literacy for safe deployment
| Citations per year | 7.33 |
|---|---|
| Citation span | 2023 - 2026 (4) |
| Highly cited | No |