AI stages, or levels, of development are defined by many people, and there may never be a real standard for the terminology. However, the following examples are common takes on the subject. See
AI and
AI in a nutshell.
1 - RULE BASED: clearcut rules. If this, do that. Expert systems were the first AI. See
expert system.
2 - CONTEXT AWARENESS: system learns from past interactions. Examples are virtual assistants. See
Siri and
Google Assistant.
3 - DOMAIN SPECIFIC: highly skilled in a specific field such as chess or Go. See
DeepMind.
4 - REASONING: resembles human intelligence. See
ChatGPT.
5 - GENERAL INTELLIGENCE: self-awareness. See
AGI.
6 - SUPER AI: outperforms humans in many areas. See
AGI.
7 - SINGULARITY: computers make all the decisions. See
technology singularity.
AI Uncovered Channel
The following is excerpted from "The 10 Stages of AI," a thought-provoking YouTube video.
1 - RULE BASED: if this, do that.
2 - CONTEXT BASED: recommendation systems.
3 - NARROW/WEAK AI: outperforms humans in single domain like playing chess.
4 - REASONING AI: solves problems on its own.
5 - ARTIFICIAL GENERAL INTELLIGENCE (AGI): can replace most humans.
6 - SUPERINTELLIGENT AI: improves itself.
7 - SELF-AWARE AI: a sense of identity.
8 - TRANSCENDANT AI: new digital and biological life forms.
9 - COSMIC AI: interstellar exploration and colonization.
10 - GODLIKE AI: rewrite reality.
Fewer Levels
AI is so overwhelmingly popular that there are many opinions.
0 - NARROW, NON-AI.
1 - EMERGING, NARROW AI
2 - COMPETENT NARROW AI equal to 50th percentile of skilled adults.
3 - EXPERT NARROW AI equal to 90th percentile of skilled adults.
4 - VIRTUOSO NARROW AI equal to 99th percentile of skilled adults.
The following was published in 2021 by Bernard Marr on LinkedIn. The explanations are excerpts.
1 - REACTIVE AI: recommendation engines, spam filters and chess playing.
2 - LIMITED MEMORY AI: historical data combined with programming to make predictions.
3 - THEORY OF MIND AI: meaningful conversation with an AI.
4 - SELF-AWARE AI: theoretical.
Another Version
The following from Harvey Ross in "AI in Plain English." The explanations are excerpts.
1 - SCRIPTED INTELLIGENCE: follow rules like software.
2 - REACTIVE AI: navigates environment in real time like robotic vacuum cleaners.
3 - LIMITED MEMORY AI: learns from limited historical data; a product recommendation.
4 - CONTEXT AWARE AI: learns from huge amounts of data like GPT models (see
ChatGPT).
5 - THEORY OF MIND AI: robotic systems anticipate human needs.
6 - SENTIENT OR SELF-AWARE AI: makes autonomous decisions.
7 - SUPER INTELLIGENT AI: surpasses human intelligence.
8 - IMMORTAL AI: theoretical.