The AI effect is a recurring phenomenon in which advances in artificial intelligence lead to a redefinition of what is considered intelligence, such that capabilities achieved by AI systems are no longer regarded as examples of "real" intelligence. As machines master tasks once thought to require human intellect, those tasks are typically reclassified as routine computation or absorbed into other domains, leaving the label of "intelligence" for problems that remain unsolved. The concept has been used to describe both a cognitive tendency and a sociotechnical pattern that shapes public perception and the history of the field.
Historian Pamela McCorduck described this as an "odd paradox" in her 2004 book Machines Who Think, noting that once a problem is solved, it is no longer considered evidence of intelligence. Researcher Rodney Brooks similarly observed in 2002 that once systems are understood, they are often regarded as "just computation." This tendency has led to the popular formulation, sometimes attributed to Tesler's theorem, that "AI is whatever hasn't been done yet."
Definition
The AI effect refers to the shift in how intelligence is defined as machines acquire new capabilities. Tasks such as playing chess, recognizing speech, or interpreting images were historically considered indicators of intelligence, but after successful automation they are often reclassified as normal engineering achievements. The phenomenon is frequently interpreted as an instance of moving the goalposts, where the threshold for what counts as "real" intelligence is continually adjusted to exclude what machines can already do.
This reclassification affects not only public discourse but also scientific terminology. When a problem is not yet fully formalised, researchers may describe it using models involving human computation, such as human-assisted Turing machines, highlighting the conceptual boundary between human and machine abilities. The AI effect thus creates a moving target for the field, as each success diminishes the perceived importance of the achievement.
Historical Examples
Game Playing
Early AI systems capable of playing games such as checkers and chess were initially regarded as demonstrations of machine intelligence. However, as these systems improved and became better understood, their achievements were often reinterpreted as examples of computation rather than intelligence. The victory of IBM's Deep Blue over Garry Kasparov in 1997 is a frequently cited example, with critics arguing that the system relied on brute-force methods rather than genuine understanding. Similar patterns accompanied earlier checkers programs, which were celebrated as milestones before being dismissed as mere algorithm execution.
Pattern Recognition
Technologies such as optical character recognition and speech recognition were once considered core problems in artificial intelligence. As these systems became reliable and widely deployed, they were increasingly treated as standard engineering solutions rather than demonstrations of machine intelligence. The transition reflects a broader pattern where successful AI techniques are assimilated into pattern recognition applications and other fields, losing their association with the AI label.
Integration into Applications
Many techniques originally developed within AI research have been incorporated into broader technological systems, including marketing, automation, and software applications. Michael Swaine reported in 2007 that AI advances are often presented as developments in other fields. Marvin Minsky observed that successful AI innovations often evolve into separate disciplines, while Nick Bostrom noted in 2006 that widely adopted technologies are often no longer labeled as AI. This integration contributes to the perception that AI is perpetually a future promise rather than a current reality.
Contemporary Discussion
The AI effect continues to be discussed in the context of recent advances in machine learning, particularly large language models and other generative AI systems. As these systems have become more widely used, some researchers and commentators have noted that their capabilities are frequently described as statistical or mechanical once understood, rather than as intelligence. For instance, different combinations of human and AI action in workflows can be considered "augmented intelligence," and specific weights within various combinations of augmentation should be used to define ethical and practical considerations contextually for those workflows.
A 2016 survey of artificial intelligence also noted that AI systems are increasingly embedded in everyday applications, reinforcing earlier observations that successful AI technologies tend to become normalized and no longer identified as AI. At the same time, the widespread commercial use of artificial intelligence has led to greater visibility of the field, contrasting with earlier periods in which AI techniques were often present but unacknowledged. This duality - simultaneous normalization and prominence - complicates simple narratives about the AI effect.
Interpretations
Cognitive Bias
Some authors describe the AI effect as a cognitive bias in which expectations of intelligence shift as machines achieve new capabilities. Under this view, humans are prone to discounting the intelligence of any system whose inner workings are known, a tendency that may stem from a perceived need to preserve human uniqueness.
Sociotechnical Perspective
Another interpretation emphasizes how technologies are reclassified over time as they become widespread and commercially successful. This perspective highlights the role of economic factors, institutional practices, and public discourse in shaping what is labeled as AI, suggesting that the AI effect is not merely a psychological quirk but a feature of how technological fields evolve.
Philosophical Debate
Some philosophers argue that reclassification reflects genuine conceptual distinctions rather than bias. They contend that as systems become formalized and predictable, it becomes appropriate to treat them as computation, reserving "intelligence" for capacities that remain mysterious or ill-defined. This view ties the AI effect to broader questions in the philosophy of mind, such as the Chinese room thought experiment and functionalist accounts of mental states.
Historical Context
During periods such as the AI winter, researchers sometimes avoided the term "artificial intelligence" due to negative perceptions and funding difficulties. This avoidance contributed to the AI effect, as successful work was often rebranded under other names to escape stigma. In the 21st century, however, the term "AI" has become widely used in public discourse and marketing, partly driven by the commercial success of companies like OpenAI, Anthropic, and Google DeepMind. This shift has made the AI effect more visible, as public expectations and definitions of intelligence are continually contested.
Broader Implications
The AI effect has been linked to broader questions about human uniqueness and the nature of intelligence. Michael Kearns suggested that people may seek to preserve a special role for humans, motivating the constant redefinition of intelligence to exclude machine achievements. Similar patterns have been observed in studies of animal cognition, where once a behavior is explained mechanistically, it is often deemed less cognitively sophisticated. Herbert A. Simon noted that artificial intelligence can provoke strong emotional reactions, reflecting the deep stakes involved in questions of intelligence and agency.
The phenomenon also has practical consequences for AI research and policy. It can affect how research priorities are set, how achievements are communicated, and how public understanding of AI develops. Recognizing the AI effect can help researchers anticipate skepticism and communicate the significance of their work more effectively.
Related Phenomena
The AI effect is closely related to several other concepts. The ELIZA effect describes the tendency to attribute more understanding to a system than it possesses, which can be seen as the inverse of the AI effect. Moravec's paradox notes that high-level reasoning requires little computation, while low-level sensorimotor skills require enormous computational resources, complicating intuitions about which tasks are "hard." The philosophy of mind offers additional frameworks, such as the Chinese room and functionalism, that bear on what counts as intelligence. In theological contexts, the "god of the gaps" argument similarly redefines the unexplained as the domain of the divine, echoing the moving goalposts pattern. These connections suggest that the AI effect is one manifestation of broader cognitive and cultural tendencies.
See Also
- ELIZA effect
- Moravec's paradox
- Chinese room
- Functionalism (philosophy of mind)
- Computational intelligence
- Artificial intelligence in video games
- God of the gaps
- Organoid intelligence