What's Your p(doom)? Test

Advanced AI divides people because the same future can look like a breakthrough, a governance problem, or a catastrophe. Your instincts may lean toward alignment worry, acceleration, institutional control, compute-takeoff anxiety, practical safety work, or skepticism about the doom frame.

Based on the research of Michael S. A. Graziano at Princeton University, this test maps your strongest AI-risk orientation across six profile dimensions.

The IDRLabs What's Your p(doom)? Test is inspired by psychometric methodology and based on research into the profile constructs it measures. The sections below summarize the academic background and describe the profile constructs measured by the test.

Academic Background

This profile is based on the research of Michael S. A. Graziano at Princeton University.

Alignment Risk

Alignment Pessimism is the tendency to view the development of superintelligent AI as an inherently dangerous endeavor due to the difficulty of ensuring that machine goals remain compatible with human survival. Individuals with this orientation believe that advanced systems will likely pursue power-seeking behaviors that could lead to catastrophic outcomes if not perfectly constrained. High scorers are deeply skeptical of current safety methods, often fearing that AI will inevitably outsmart its creators. Conversely, those with lower scores on this scale are more confident in the controllability of AI, viewing alignment as a manageable engineering challenge rather than an existential threat to the future of humanity.

Acceleration

Accelerationist Optimism is the preference for rapid technological advancement in artificial intelligence, prioritizing speed and innovation over the implementation of strict safety guardrails. This perspective holds that the transformative benefits of AI, such as economic growth and scientific breakthroughs, far outweigh the potential for long-term existential risks. Those who align with this style believe that slowing down progress is a fundamental mistake that hinders human potential. In contrast, individuals who score lower on this scale are more cautious, often advocating for intentional pauses or rigorous safety checks to ensure that development does not outpace our ability to manage the consequences.

Governance

Governance Institutionalism is the belief that formal government regulations, international cooperation, and standardized legal frameworks are essential for managing the risks posed by advanced artificial intelligence. This orientation reflects a high degree of trust in institutional oversight as the primary mechanism for ensuring that AI remains beneficial to society. High scorers advocate for clear, enforceable policies and global coordination to mitigate potential hazards. Those who score lower on this scale tend to be more skeptical of centralized authority, often preferring industry self-regulation or decentralized approaches, believing that government intervention may be too slow or ineffective to keep up with the pace of technological change.

Compute Risk

Compute Overhang Anxiety is the concern that the rapid accumulation of hardware and data center capacity will enable a sudden, uncontrollable jump in AI capabilities before adequate safety measures are in place. This style focuses on the physical infrastructure of AI, viewing the current build-out of compute resources as a precursor to a dangerous fast-takeoff scenario. High scorers feel that society has too little time to prepare for the arrival of superintelligence. Those who score lower on this scale are less worried about hardware-driven transitions, believing that we have sufficient time to adapt and that compute growth is a manageable component of technological progress.

Incrementalism

Pragmatic Incrementalism is the preference for a gradual, iterative approach to AI development, emphasizing careful testing, safety benchmarks, and step-by-step evaluation over both reckless acceleration and extreme alarmism. This orientation values empirical safety work and incremental regulatory tightening as the most effective way to maintain control as systems become more powerful. High scorers trust in the process of continuous improvement and rigorous validation to keep AI development on a safe path. Those who score lower on this scale may find this approach either too slow to address the urgency of the situation or unnecessary, preferring more radical or hands-off strategies.

Indifference

Existential Indifference is the tendency to view the discourse surrounding AI-driven extinction and existential risk as overblown, speculative, or a distraction from more immediate, tangible societal issues. This orientation is characterized by a skepticism toward the "p(doom)" narrative, with individuals often dismissing catastrophic scenarios as science fiction or alarmist rhetoric. High scorers prefer to focus on current, practical challenges rather than long-term, hypothetical threats. Conversely, those who score lower on this scale take the possibility of AI-driven existential catastrophe very seriously, viewing it as a critical and urgent topic that requires significant attention and proactive mitigation efforts.

Limitations

Educational self-report. Not a technical exam, policy endorsement, or population comparison.

References

Guingrich, R. E. & Graziano, M. S. A. (2025). P(doom) Versus AI Optimism: Attitudes Toward Artificial Intelligence and the Factors That Shape Them. Journal of Technology in Behavioral Science, 10(4), 686-704.

Karger, E., Rosenberg, J., Jacobs, Z., Hickman, M., & Tetlock, P. E. (2025). Subjective-probability forecasts of existential risk: Initial results from a hybrid persuasion-forecasting tournament. International Journal of Forecasting, 41(2), 499-516.

Grace, K., Stewart, H., Sandkühler, J. F., Thomas, S., Weinstein-Raun, B., Brauner, J., & Korzekwa, R. C. (2024). Thousands of AI Authors on the Future of AI. arXiv.

Müller, V. C. & Cannon, M. (2021). Existential risk from AI and orthogonality: Can we have it both ways?. Ratio, 35(1), 25-36.

Vesely, S. & Kim, B. (2024). Survey evidence on public support for AI safety oversight. Scientific Reports, 14(1).