Systemic Risks
Governing General Purpose AI: A Comprehensive Map of Unreliability, Misuse and Systemic Risks
Maham & Küspert (2023)
AI-driven concentration of power and resources within certain entities or groups, especially those with access to or ownership of powerful AI systems, leading to inequitable distribution of benefits and increased societal inequality.
"In addition to risks stemming from the unreliability or misuse of general purpose AI models, further Systemic Risks can originate from the centralisation of general purpose AI development as well as the rapid integration of these models into our lives."(p. 17)
Sub-categories (3)
Economic Power Centralisation and Inequality
"Increasingly advanced general purpose AI models pose the risk of a concentration of economic power and exacerbation of existing inequalities through disparities in effective access to these models. This can materialise on multiple levels, between developers of general purpose AI models and companies building applications on them, between individuals and between countries on a global scale."
6.1 Power centralization and unfair distribution of benefitsIdeological Homogenization from Value Embedding
"The increasing integration of general purpose AI models into every-day life raises concerns around their embedded normative values. The reach of a small number of AI models to a large number of people around the world can make these value judgements unprecedently impactful, potentially leading to increased ideological homogenization."
1.3 Unequal performance across groupsDisruptions from Outpaced Societal Adaptation
"Although the implementation of general purpose AI models as automation tools could be a major opportunity, overly rapid adoption of this technology at scale might outpace the ability of society to adapt effectively. This could lead to a variety of disruptions, including challenges in the labour market, the education system and public discourse, and various mental health concerns."
5.2 Loss of human agency and autonomyOther risks from Maham & Küspert (2023) (10)
Misuse Risks
4.0 Malicious Actors & MisuseMisuse Risks > Cybercrime
4.3 Fraud, scams, and targeted manipulationMisuse Risks > Biosecurity Threats
4.2 Cyberattacks, weapon development or use, and mass harmMisuse Risks > Politically motivated misuse
4.1 Disinformation, surveillance, and influence at scaleDiscrimination and Stereotype Reproduction
1.1 Unfair discrimination and misrepresentationMisinformation and Privacy Violations
3.1 False or misleading information