What are common AI Threat Actors?

• A. Insider threats, nation states, cybercriminals, and AI developers
• B. End users
• C. Regulators
• D. Investors

Answer: (A)

In AI security, the most relevant threat actors are insiders, nation-states, cybercriminals, and AI developers. Insider threats involve people within an organization who have legitimate access to data, models, or systems and may misuse or exfiltrate information, sabotage processes, or bypass controls. Nation-states act as threat actors when they pursue strategic advantages through espionage, disruption, or manipulation of AI capabilities, targeting labs, data pipelines, or deployed models. Cybercriminals are driven by financial gain and may steal data, steal or misuse models, or deploy ransomware against AI infrastructure, or attempt data poisoning to degrade performance. AI developers can pose risks as creators and maintainers of AI systems; they might introduce backdoors, hidden vulnerabilities, or lax controls either inadvertently or intentionally, and supply chain involvement (through vendors or collaborators) can amplify these risks.

These categories capture the primary motive and capability profiles seen in AI-related risk scenarios. End users can pose risk in certain contexts but aren’t typically treated as a distinct threat-actor category in AI security; regulators and investors are external stakeholders focused on oversight and funding, not attackers.