As artificial intelligence continues to permeate education and training, the language we use to describe these systems requires careful scrutiny. Two recent articles highlight how the vocabulary surrounding AI is often misleading, obscuring the realities of the technology and shifting accountability away from its creators. In educational settings, where trust and clarity are paramount, understanding the implications of these words is a critical component of digital literacy.

The first piece, a newsletter post by Dr. Sam Illingworth and The Strategic Linguist, argues that the everyday vocabulary of AI misleads by borrowing terms from the human mind. The authors suggest that words like "intelligence" and "hallucination" are too flattering to machines and imply a level of understanding that simply does not exist. They propose a new dictionary, replacing "intelligence" with "prediction" and "hallucination" with "fabrication," to better reflect the statistical pattern matching and generative acts these systems actually perform .

This perspective aligns with the arguments presented by Lucy Suchman in her commentary for Big Data & Society. Suchman challenges the "uncontroversial 'thingness' of AI," urging critical scholars to question the rhetorical moves that position AI as a stable, agential entity. She emphasizes that the nominalization of AI mystifies its agency and obscures the material practices and political economies that underwrite it. By treating AI as a self-evident actor, we risk missing the opportunity to trace its sources of power and demystify its referents .

In the context of education, the implications of these linguistic choices are profound. When we refer to a system as "intelligent," we may inadvertently extend to it the trust we reserve for human educators who actually understand the material. As Illingworth and The Strategic Linguist note, the word "launders a guess into a judgement." Similarly, describing a system's errors as "hallucinations" medicalizes the failure, suggesting a temporary glitch in a reliable mind rather than a fundamental characteristic of the technology. This framing protects the companies that build and deploy these systems by erasing their responsibility from the narrative .

Suchman's call to destabilize the figure of AI is particularly relevant for educators and administrators evaluating these tools. She warns against the strategic vagueness of the term "AI," which serves the interests of its promoters by allowing it to act as a floating signifier. When AI is invoked as a singular and autonomous agent, it becomes difficult to locate invested actors and specify relevant classes of technology. This makes it challenging to hold developers accountable for the biases and limitations inherent in their systems .

To foster a more critical engagement with AI in vocational education and training, we must adopt a vocabulary that accurately describes what these technologies do. Illingworth and The Strategic Linguist offer several replacements: using "prediction" instead of "intelligence," "fabrication" instead of "hallucination," asking "general at what?" instead of referencing Artificial General Intelligence (AGI), using "mimicry" instead of "consciousness," and referring to an "operator" rather than an "agent" .

By changing our language, we can shift the focus from the imagined capabilities of the machine to the tangible actions of the people who build, deploy, and use these systems. This is not mere pedantry; as the authors of the newsletter point out, "words are policy." In educational environments, where the goal is to equip learners with the skills and critical thinking necessary for the modern workforce, the words we choose to describe our tools shape how we understand, regulate, and integrate them into our classrooms. A critical AI literacy must begin with the words we use.

References

[1] Illingworth, S., & The Strategic Linguist. (2026). We Are Using the Wrong Words for AI. Slow AI.

[2] Suchman, L. (2023 ). The uncontroversial ‘thingness’ of AI. Big Data & Society, 10(2).