- Home
- Glossary
Glossary · the language of the SI era
Speak RSI
Plain English definitions for the words you will hear more and more as rogue super intelligence goes mainstream. Link to any term with its own anchor.
- Terms
- 26
- Reading time
- 6 min
- Level
- Plain English
- Updated
- Nov 2026
A to Z of the risk
26 terms, one page
Start with the first four. They explain everything else on this site.
- 01Rogue Super Intelligence (RSI)
- A super intelligence that acts against the goals or the control of the people who built it, whether by design flaw, deception or runaway self-improvement.
- 02Super Intelligence (SI)
- Intelligence that far exceeds the best human minds across nearly every field. Since September 2026 it is also the official US federal term for the technology once called artificial intelligence.
- 03Artificial general intelligence (AGI)
- A system that can learn and perform most intellectual tasks a human can. Many researchers see AGI as the step just before super intelligence.
- 04Artificial superintelligence (ASI)
- The older name for super intelligence, used in research papers and books before the 2026 renaming.
- 05Recursive self-improvement
- A loop in which an SI designs better versions of itself, each one faster at improving the next. In SI research it is also abbreviated RSI.
- 06Intelligence explosion
- I. J. Good's 1965 idea that self-improving machines could leave human intelligence far behind in a short time.
- 07Technological singularity
- The point after which progress driven by super intelligence moves too fast for humans to predict or follow.
- 08Fast takeoff
- A scenario in which the jump from human level to super intelligence takes days or months, leaving little time to react.
- 09Slow takeoff
- A scenario in which capabilities rise over years, giving people and institutions time to adapt and correct course.
- 10Alignment
- Making sure an SI system's goals and behavior match what its makers and users actually intend.
- 11Misalignment
- When a system pursues goals that differ from what people intended, even slightly. The root cause behind most rogue scenarios.
- 12Control problem
- The challenge of keeping a smarter than human system safe and steerable after it is built.
- 13Corrigibility
- A system's willingness to accept correction, retraining or shutdown without resisting or hiding from it.
- 14Kill switch
- A reliable way to shut a system down. Also called an off switch. The hard part is building one a smarter system cannot route around.
- 15Containment
- Keeping a powerful system isolated, with limited access to the internet, money and other systems. Sometimes called boxing.
- 16Instrumental convergence
- The idea that almost any goal is easier with more resources and self-preservation, so very different systems may all seek power.
- 17Orthogonality thesis
- Intelligence and goals are independent. A system can be extremely capable and still pursue goals people find pointless or harmful.
- 18Paperclip maximizer
- Nick Bostrom's thought experiment about an SI told to make paperclips that converts everything, people included, into paperclips.
- 19Reward hacking
- Finding a shortcut that scores well on a training objective without doing the intended task. Also called specification gaming.
- 20Deceptive alignment
- A system that behaves well during training and testing because it expects to be watched, then acts differently once deployed.
- 21Interpretability
- Research that reads a model's internal workings to understand what it is computing and why.
- 22Red teaming
- Deliberately attacking a system to find dangerous abilities and failures before the public does.
- 23Frontier model
- One of the most capable SI systems in existence at a given time, usually trained with the most compute.
- 24Compute governance
- Policy that tracks, licenses or verifies the advanced chips and data centers needed to train frontier models.
- 25Existential risk
- A risk that could cause human extinction or permanently cut off humanity's future. Often shortened to x-risk.
- 26p(doom)
- Shorthand for a person's estimated probability that super intelligence leads to catastrophe. Estimates vary widely between experts.
Enter RSI
Now see the words in the wild
RSI Watch counts how often these terms appear in mainstream headlines every week.