comprehensive AI services
In plain wordsThe idea that superintelligent-level AI could arrive as many specialized AI tools, each doing a bounded job, rather than as one all-powerful AI agent with goals of its own.
A model of advanced AI in which general, even superintelligent, capability is delivered by many bounded AI services, each performing a specific task with limited resources in limited time, rather than by unified agents pursuing open-ended goals. Its author argued that AI development is already organized as research and engineering pipelines producing task-focused systems, that automating AI research would make new services available on demand, and that this could deliver the benefits usually attributed to AGIartificial general intelligence, n.: AI that could learn and do almost any mental task a capable adult can, rather than excelling at one narrow job. Experts disagree on the exact bar and on when, or whether, it will arrive. without building a goal-directed agent whose motivations must be aligned. CAIS shifted part of AI-safety discussion from controlling one superintelligent agent toward governing a world of powerful tools. Critics replied that economic and military pressures favor integrating services into agents, that coordinated services can themselves amount to an agent, and that the model does not address misuse; the turn toward agentic AIagentic AI, n.: AI built to pursue goals on its own — planning, taking actions and adjusting along the way — instead of only responding to one prompt at a time. 'Agentic' describes how much of this independence a system has. since 2024 is often cited in support of these objections.
CAIS imagined superintelligence as a toolbox; the market built agents instead.
Origin
Proposed by K. Eric Drexler, the nanotechnology pioneer, in the Future of Humanity Institute technical report Reframing Superintelligence: Comprehensive AI Services as General Intelligence (2019).
Related
- tool AItool AI, n.: AI built to be used like an instrument, answering, suggesting or calculating when asked, rather than pursuing goals on its own. Some argue that this is the only kind we should build.
- oracle AIoracle AI, n.: An AI kept to answering questions, without acting in the world on its own. Proposed as a safer kind of superintelligence, though its answers could still steer the people who read them.
- agentic AIagentic AI, n.: AI built to pursue goals on its own — planning, taking actions and adjusting along the way — instead of only responding to one prompt at a time. 'Agentic' describes how much of this independence a system has.
- AI agentAI agent, n.: An AI system that does not just answer but acts: given a goal, it decides on steps, uses tools such as a browser or code editor, checks the results and keeps going until the job is done.
- automated AI researchautomated AI research, n.: Using AI systems to do the work of AI researchers, such as running experiments and writing code for new models, so that AI helps build its own successors.
- superintelligencesuperintelligence, n.: A mind, usually imagined as an AI, that would leave even the cleverest people far behind in nearly every kind of thinking, not just in one game or skill. None is known to exist.
- AI agentAI agent, n.: An AI system that does not just answer but acts: given a goal, it decides on steps, uses tools such as a browser or code editor, checks the results and keeps going until the job is done.
Sources
- Drexler, K. E. (2019). Reframing Superintelligence: Comprehensive AI Services as General Intelligence. Technical Report 2019-1, Future of Humanity Institute, University of Oxford.
Mentioned in
oracle AIoracle AI, n.: An AI kept to answering questions, without acting in the world on its own. Proposed as a safer kind of superintelligence, though its answers could still steer the people who read them., tool AItool AI, n.: AI built to be used like an instrument, answering, suggesting or calculating when asked, rather than pursuing goals on its own. Some argue that this is the only kind we should build.