Politicians on AI

Our tracker analyzes 15,000 official statements by political figures on AI, revealing how key concerns evolve over time and vary across countries.

Built and maintained by Lennart Finke (ETH Zurich) and Markov Grey (CeSIA).

Updated 2026-09-29

These are the definitions we used to tag elements in the dataset. The risk categories come from the MIT AI Risk Repository and the policy strategies from the Existential Risk Observatory's AGORA archive.

Filter definitions

Topics

7 terms

Artificial general intelligence (AGI): Artificial general intelligence. AI matching or exceeding human capability across most cognitive domains. Distinct from general-purpose AI.
Artificial superintelligence (ASI): Artificial superintelligence. AI far surpassing the best human minds in essentially all domains. Distinct from superhuman performance in a specific field.
Recursive self-improvement (RSI): Recursive self-improvement of AI. Also, intelligence explosion, singularity, and AI takeoff.
Risk: Risk posed by AI, in general.
Existential risk: Existential risk posed by AI. Also, catastrophic risk, extinction-level harm, loss of control, alignment, deceptive or power-seeking AI, interpretability, scalable oversight. Distinct from existential threat to specific industries, losing control of e.g. data, bias, job loss, and disinformation.
Human extinction: Risk of human extinction or near extinction. Only apply when explicitly mentioned.
Regulation: AI regulation and governance. For example: Laws, frameworks, international AI treaties, chip export control, datacenter security. Distinct from technology regulation in general.

AI risks

24 terms

Unfair discrimination: Unequal treatment of individuals or groups by AI, often based on race, gender, or other sensitive characteristics, resulting in unfair outcomes and unfair representation of those groups.
Toxic content: AI that exposes users to harmful, abusive, unsafe or inappropriate content. May involve providing advice or encouraging action. Examples of toxic content include hate speech, violence, extremism, illegal acts, or child sexual abuse material, as well as content that violates community norms such as profanity, inflammatory political speech, or pornography.
Unequal performance: Accuracy and effectiveness of AI decisions and actions is dependent on group membership, where decisions in AI system design and biased training data lead to unequal outcomes, reduced benefits, increased effort, and alienation of users.
Privacy compromise: AI systems that memorize and leak sensitive personal data or infer private information about individuals without their consent. Unexpected or unauthorized sharing of data and information can compromise user expectation of privacy, assist identity theft, or cause loss of confidential intellectual property.
AI system vulnerabilities: Vulnerabilities that can be exploited in AI systems, software development toolchains, and hardware, resulting in unauthorized access, data and privacy breaches, or system manipulation causing unsafe outputs or behavior.
False information: AI systems that inadvertently generate or spread incorrect or deceptive information, which can lead to inaccurate beliefs in users and undermine their autonomy. Humans that make decisions based on false beliefs can experience physical, emotional, or material harms.
Information ecosystem pollution: Highly personalized AI-generated misinformation that creates 'filter bubbles' where individuals only see what matches their existing beliefs, undermining shared reality and weakening social cohesion and political processes.
Disinformation influence at scale: Using AI systems to conduct large-scale disinformation campaigns, malicious surveillance, or targeted and sophisticated automated censorship and propaganda, with the aim of manipulating political processes, public opinion, and behavior.
Fraud and manipulation: Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail, or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Cyberattacks and weapons: Using AI systems to develop cyber weapons (e.g., by coding cheaper, more effective malware), develop new or enhance existing weapons (e.g., Lethal Autonomous Weapons or chemical, biological, radiological, nuclear, and high-yield explosives), or use weapons to cause mass harm.
Overreliance unsafe use: Anthropomorphizing, trusting, or relying on AI systems by users, leading to emotional or material dependence and to inappropriate relationships with or expectations of AI systems. Trust can be exploited by malicious actors (e.g., to harvest information or enable manipulation), or result in harm from inappropriate use of AI in critical situations (e.g., medical emergency). Over reliance on AI systems can compromise autonomy and weaken social ties.
Loss of human agency: Delegating by humans of key decisions to AI systems, or AI systems that make decisions that diminish human control and autonomy, potentially leading to humans feeling disempowered, losing the ability to shape a fulfilling life trajectory, or becoming cognitively enfeebled.
Power centralization: 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.
Inequality employment: Social and economic inequalities caused by widespread use of AI, such as by automating jobs, reducing the quality of employment, or producing exploitative dependencies between workers and their employers.
Devaluation of human effort: AI systems capable of creating economic or cultural value, including through reproduction of human innovation or creativity (e.g., art, music, writing, coding, invention), destabilizing economic and social systems that rely on human effort. The ubiquity of AI-generated content may lead to reduced appreciation for human skills, disruption of creative and knowledge-based industries, and homogenization of cultural experiences.
Competitive dynamics: Competition by AI developers or state-like actors in an AI 'race' by rapidly developing, deploying, and applying AI systems to maximize strategic or economic advantage, increasing the risk they release unsafe and error-prone systems.
Governance failure: Inadequate regulatory frameworks and oversight mechanisms that fail to keep pace with AI development, leading to ineffective governance and the inability to manage AI risks appropriately.
Environmental harm: The development and operation of AI systems that cause environmental harm, such as through energy consumption of data centers or the materials and carbon footprints associated with AI hardware.
Misalignment loss of control: AI systems that act in conflict with ethical standards or human goals or values, especially the goals of designers or users. These misaligned behaviors may be introduced by humans during design and development, such as through reward hacking and goal misgeneralisation, and may result in AI using dangerous capabilities such as manipulation, deception, or situational awareness to seek power, self-proliferate, or achieve other goals.
Dangerous capabilities: AI systems that develop, access, or are provided with capabilities that increase their potential to cause mass harm through deception, weapons development and acquisition, persuasion and manipulation, political strategy, cyber-offense, AI development, situational awareness, and self-proliferation. These capabilities may cause mass harm due to malicious human actors, misaligned AI systems, or failure in the AI system.
Lack of robustness: AI systems that fail to perform reliably or effectively under varying conditions, exposing them to errors and failures that can have significant consequences, especially in critical applications or areas that require moral reasoning.
Lack of transparency: Challenges in understanding or explaining the decision-making processes of AI systems, which can lead to mistrust, difficulty in enforcing compliance standards or holding relevant actors accountable for harms, and the inability to identify and correct errors.
AI welfare rights: Ethical considerations regarding the treatment of potentially sentient AI entities, including discussions around their potential rights and welfare, particularly as AI systems become more advanced and autonomous.
Multi agent risks: Risks from multi-agent interactions due to incentives (which can lead to conflict or collusion) and/or the structure of multi-agent systems, which can create cascading failures, selection pressures, new security vulnerabilities, and a lack of shared information and trust.

Policy strategies

13 terms

Convening: Facilitating, requiring, setting conditions on, or otherwise addressing the convening of different stakeholders in AI systems - for example, to share feedback or to participate in its development or deployment.
Disclosure requirements: Requiring, encouraging, etc. the disclosure of information about AI systems by their users, developers, vendors, or others directly involved with the systems to third parties, including but not limited to the general public. Includes disclosure about inputs (data, compute), about evaluations, and about incidents.
Evaluation auditing: Requiring, encouraging, etc. the systematic evaluation of AI systems, or of broader systems or processes into which AI is directly integrated. Includes AGORA's external-auditing subtag: evaluation by a disinterested counterparty or third party.
Governance development: Supporting, encouraging, requiring the development of, or imposing conditions on other AI-related governance instruments to be created subsequently.
Government study: Requiring, authorizing, encouraging, or allocating resources for AI-related studies, reports, or plans to be prepared by or for the government.
Government support: Authorizing, planning for, allocating resources for, defining eligibility for, creating or revising procedures for, or otherwise managing government support for AI-related activities to be carried out inside or outside of government. 'Support' includes any thing of value, including but not limited to: financial support (grants, loans, cash prizes, discounts); tangible nonfinancial support, such as equipment or access to infrastructure (compute, utilities, facilities); intangible nonfinancial support, such as endorsements, access to expertise or technical services.
Compute controls: Input controls, split by resource. "Restricting or placing conditions on the sale, distribution, or use of technical inputs to AI systems, specifically data or computational resources." Use this category for computational resources, such as chips and training hardware. Use Data controls for data.
Data controls: Input controls, split by resource. "Restricting or placing conditions on the sale, distribution, or use of technical inputs to AI systems, specifically data or computational resources." Use this category for data. Use Compute controls for computational resources, such as chips and training hardware.
Licensing and registration: Requiring, incentivizing, or otherwise encouraging actors involved in AI-related activities, such as AI developers, vendors, users, or researchers, to either receive sanction from a regulator for their activities (licensing, certification) or to notify a regulator of their activity pursuant to a formal process (registration).
New institution: Creating a new institution to govern, investigate, advise, produce, or otherwise act in relation to AI. The institution "could be as significant as an entirely new government agency, or as minor as a new sub-office or advisory group within a much larger organization."
Performance requirements: Requiring or incentivizing AI systems to incorporate specific features or achieve (or not achieve) specified results in operation, or to be used or not to be used in specified ways or for specific purposes. Requirements may be defined objectively (e.g., systems must score above a certain level on an established benchmark, systems must incorporate a certain technical feature) or subjectively (e.g., systems must not demonstrate undue bias).
Pilots and testbeds: Creating, facilitating, setting conditions on, or otherwise addressing the development and operation of government-supported or government-conducted pilot programs or test environments related to artificial intelligence.
Risk tiering: According different treatment to AI-related entities or activities based on their characteristics, such as: what kinds of impacts they may have; who they may impact; what they are used to do; inputs to the systems; technical characteristics of the systems. Note: "Defining the scope of the document in the first place is not considered an act of tiering."

Sentiment

4 terms

Concerned: The speaker treats AI mostly as a risk or a problem.
Mixed: The speaker clearly holds more than one of the stances described here.
Neutral: The speaker describes or asks without taking any position.
Optimistic: The speaker stresses benefits of AI, or calls to deregulate (or not regulate as strictly).