Skip to main content
MIT AI Risk Initiative
BackMalicious Use (Intentional)
Home/Risks/Hendrycks, Mazzeika & Woodside (2023)/Malicious Use (Intentional)
Model prediction uncertainty
Bioterrorism

Malicious Use (Intentional)

An Overview of Catastrophic AI Risks

Hendrycks, Mazzeika & Woodside (2023)

SourceDOI
Category

"empowering malicious actors to cause widespread harm"(p. 6)

Entity— Who or what caused the harm

Human

Due to a decision or action made by humans

AI system

Due to a decision or action made by an AI system

Other

Due to some other reason or is ambiguous

Intent— Whether the harm was intentional or accidental

Intentional

Due to an expected outcome from pursuing a goal

Unintentional

Due to an unexpected outcome from pursuing a goal

Other

Without clearly specifying the intentionality

Timing— Whether the risk is pre- or post-deployment

Pre-deployment

Occurring before the AI is deployed

Post-deployment

Occurring after the AI model has been trained and deployed

Other

Without a clearly specified time of occurrence

Sub-categories (4)

Bioterrorism

"AIs with knowledge of bioengineering could facilitate the creation of novel bioweapons and lower barriers to obtaining such agents."

4.2 Cyberattacks, weapon development or use, and mass harm
AI systemIntentionalPost-deployment

Unleashing AI Agents

"people could build AIs that pursue dangerous goals’"

4.2 Cyberattacks, weapon development or use, and mass harm
HumanIntentionalPre-deployment

Persuasive AIs

"The deliberate propagation of disinformation is already a serious issue, reducing our shared understanding of reality and polarizing opinions. AIs could be used to severely exacerbate this problem by generating personalized disinformation on a larger scale than before. Additionally, as AIs become better at predicting and nudging our behavior, they will become more capable at manipulating us"

4.1 Disinformation, surveillance, and influence at scale
AI systemOtherPost-deployment

Concentration of Power

"Governments might pursue intense surveillance and seek to keep AIs in the hands of a trusted minority. This reaction, however, could easily become an overcorrection, paving the way for an entrenched totalitarian regime that would be locked in by the power and capacity of AIs"

6.1 Power centralization and unfair distribution of benefits
HumanIntentionalOther

Other risks from Hendrycks, Mazzeika & Woodside (2023) (13)

AI Race (Environmental/Structural)

6.4 Competitive dynamics
HumanIntentionalOther

AI Race (Environmental/Structural) > Military AI Arms Race

6.4 Competitive dynamics
HumanIntentionalOther

AI Race (Environmental/Structural) > Corporate AI Race

6.4 Competitive dynamics
HumanUnintentionalOther

Rogue AIs (Internal)

7.1 AI pursuing its own goals in conflict with human goals or values
AI systemIntentionalOther

Rogue AIs (Internal) > Proxy Gaming

7.1 AI pursuing its own goals in conflict with human goals or values
AI systemIntentionalOther

Rogue AIs (Internal) > Goal Drift

7.1 AI pursuing its own goals in conflict with human goals or values
AI systemIntentionalOther
View all 13 risks from this paper →