The researchers searched Scopus, Web of Science, and ProQuest, databases that earlier systematic reviews had not always covered fully. From an initial pool of 2,212 articles, duplicates and retracted papers were removed, and two independent coders screened the remainder with a Cohen’s Kappa agreement of 0.807, a level the authors describe as almost perfect. Full-text screening, guided by the PRISMA-ScR reporting framework, ultimately yielded 37 eligible reports. The deliberate exclusion of the keyword “voice” from the search string, the authors explain, was a methodological safeguard: voice technology is inconsistently mentioned in titles and abstracts, and broader AI-related terms maximized retrieval sensitivity.

One of the clearest patterns to emerge is the explosive growth of the field. Publication output rose steadily through the review period, with nearly a quarter of the studies appearing in 2024 and more than 40 percent in 2025 and 2026. The authors attribute this surge largely to the arrival of ChatGPT in late 2022, whose ability to maintain conversational consistency and generate contextually relevant responses made it an attractive speaking partner. Indeed, ChatGPT was the most frequently studied tool, appearing in 12 of the 19 reports on general-purpose systems, and it dominated the review’s findings.

To organize the scattered literature, the team built an analytical framework combining Richards’ classification of talk as interaction, transaction, and performance with Bibauw and colleagues’ typology of dialogue-based computer-assisted language learning systems. Under this lens, the 18 distinct voice-based AI systems identified across the studies fell into three groups: education-oriented platforms designed for language learning, general-purpose generative systems such as ChatGPT, and intelligent personal assistants like Google Assistant and Amazon Alexa. Each system played one or more of three fundamental roles: a speaking facilitator offering input and emotional support, a conversational partner for transactional or interactive exchanges, and a corrective feedback provider.

Talk as transaction proved the dominant speaking activity, reported in 30 studies. These tasks ranged from short dialog strings and role-based exchanges, such as making a hotel reservation by phone, to topic-based simulated conversations and problem-solving communicative tasks. Fifteen studies reported talk as interaction, where learners chatted casually, played guessing games, or asked everyday questions, while talk as performance, in which students delivered extended monologues followed by AI feedback, was the least common at six reports. The alignment, the authors argue, reflects how general-purpose generative AI is naturally suited to genuine information-exchange dialogue rather than mechanical drill practice.

The review’s synthesis of reported outcomes is strikingly positive, though the authors urge caution. Twenty studies documented improvements in overall speaking performance and specific linguistic components, including pronunciation, vocabulary, fluency, and grammar. Five experimental studies found that chatbot interaction facilitated meaning construction, helping learners generate contextually appropriate utterances. One cited study showed that out-of-class interaction with Google Assistant significantly enhanced learners’ oral proficiency in ways comparable to communication with native speakers. On the affective side, 26 reports described increased speaking confidence, enjoyment, motivation, autonomy, or reduced anxiety, a pattern the authors connect to self-determination theory: chatbots satisfied learners’ needs for autonomy, relatedness, and competence by offering tireless, judgment-free practice at any hour.

Yet the picture is not uniformly rosy. Twelve studies recorded negative emotions, most often stemming from the systems’ lack of human-likeness, including mechanical tone, limited emotional intelligence, and the absence of non-verbal communication. Technological complaints were even more widespread, with 17 papers reporting frustration over slow processing, problematic responses, and breakdowns caused by limitations in natural language processing or speech recognition. When communication failed, learners typically abandoned the exchange, rephrased, repeated, or code-switched. Notably, one study tracking usage over two months found that initial enthusiasm for an intelligent personal assistant dropped sharply within days, with half of the participants ceasing active use, a warning about the novelty effect that colors much of the optimistic evidence.

Methodologically, the review found the field still in an exploratory stage. Twenty-six of the 37 studies used explanatory mixed-methods designs, and 17 of those relied on short experimental interventions with pre- and post-tests on relatively small samples. More than half of the research took place in formal classroom settings under teacher supervision, leaving learners’ autonomous, out-of-class use largely unexamined, precisely where a conversational AI partner might offer the greatest advantage in EFL contexts. The authors also note that most studies measured discrete, easily quantifiable outcomes such as pronunciation and grammar, while dimensions like discourse management and interactive communication received minimal attention, and inconsistencies between self-reported feelings and objectively measured performance were rarely investigated.

These gaps shape the review’s recommendations. The authors call for longitudinal and repeated-measures studies of learners’ multidimensional engagement using digital trace data alongside self-reports, more sophisticated acceptance models that trace the transition from intention to actual sustained use, and research into how chatbots can be pedagogically integrated rather than simply compared with human practice. They also flag interdisciplinary questions about overreliance on AI, social costs of prolonged human-machine talk, environmental impacts, and educational equity for underprivileged learners who may lack both quality instruction and the devices or connectivity that voice-based AI requires.

For educators, the practical message is concrete. Teachers should provide affective, capacity, and behavior support: encouraging learners to see AI as a legitimate substitute interlocutor, training them in AI literacy and prompt design so conversations succeed, and helping them persist through communication breakdowns. Because learners still express dissatisfaction with AI’s feedback precision, the review positions human teachers as the primary scaffolders and correctors, with chatbots serving as tireless conversation partners that extend English-speaking opportunities far beyond the classroom walls.

Subject of Research: Voice-based AI chatbots supporting meaning-focused English speaking skills among EFL learners in higher education

Article Title: A scoping review of voice based AI chatbots in EFL learners’ meaning focused speaking in higher education

Article References: Phan, T. T. H., Phan, N. H. S., & Trang, L. T. (2026). A scoping review of voice based AI chatbots in EFL learners’ meaning focused speaking in higher education. Discover Education, 5(1), Article 968. https://doi.org/10.1007/s44217-026-02167-5

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02167-5

Keywords: voice-based AI, AI chatbots, EFL learners, English speaking, higher education, scoping review, ChatGPT, conversational AI, language learning, speaking anxiety, meaning-focused speaking, educational technology

The research, published in the journal Behavior Research Methods by Beatriz Bermúdez-Margaretto of the University of Salamanca and colleagues, collected valence and arousal ratings for more than 2,000 English words from over 600 Spanish-speaking college students learning English as a foreign language, or EFL. Valence captures how pleasant or unpleasant a word feels, ranging from deeply negative to strongly positive, while arousal measures the intensity of the emotional reaction it provokes, from calming to invigorating. These two dimensions form the backbone of the circumplex model of affect, one of the most influential frameworks in emotion research, and they have been systematically catalogued for English native speakers since the landmark Affective Norms for English Words dataset created by Margaret Bradley and Peter Lang in 1999.

What has been missing, the authors argue, is an equivalent resource for the enormous population of learners whose exposure to a second language is confined to formal, instructional settings. Previous research has consistently shown that bilinguals process emotional language differently from monolinguals, and that the gap between a first and second language can be striking. Physiological studies have found reduced skin-conductance responses to emotional phrases in a second language, eye-tracking experiments reveal muted pupillary reactions to affective words, and surveys of multilinguals famously report that swearwords feel blunted and clinical when spoken in a tongue learned late. Yet most existing second-language norms have been collected from immersed bilinguals, people living and studying inside the language community, whose daily experiences knit emotional associations into the new lexicon in ways that classroom learners rarely replicate.

To fill that gap, the Spanish research team recruited a large, homogeneous and representative sample of university students whose English instruction had taken place almost entirely in classrooms. Participants rated each word on nine-point scales for valence and arousal, following procedures similar to those used in earlier norming projects. The researchers also gathered proficiency data using the LexTALE lexical test and detailed language-history information through the Language Experience and Proficiency Questionnaire, allowing them to characterize a population that is statistically unusual in the bilingualism literature: late learners, embedded in a Spanish-dominant society, whose contact with English is bounded by lessons, textbooks and sporadic media consumption.

The psychometric quality of the resulting database proved robust. Reliability analyses and checks against external validity benchmarks indicated that both the valence and arousal dimensions behaved as established affective measures should, with ratings for known positive and negative words falling where theory predicts. The team then performed the analysis that gives the dataset its scientific punch: direct comparisons between the new EFL norms, existing ratings from immersed learners of English as an additional language, and native English speaker norms drawn from the massive Warriner, Kuperman and Brysbaert database of more than 13,000 English lemmas. Those comparisons revealed systematic patterns rather than random noise, indicating that the learning environment itself leaves measurable fingerprints on how emotional content is represented in the mental lexicon.

Two findings stand out. First, the discrepancy between native speaker ratings and EFL learner ratings grew systematically as a function of affective, lexical and sensory experience factors. In other words, the emotional dimensions of a word were not distorted uniformly; the divergence tracked properties such as how strongly the word engages emotion, how familiar and frequent it is, how early in life it is typically acquired, and how much it appeals to the senses. This pattern is consistent with grounded and embodied accounts of cognition, which hold that word meanings are scaffolded on sensory and motor experience. A classroom learner simply accumulates fewer embodied encounters with an English word than a native speaker does, and the deficit appears to scale with exactly those experience-dependent factors.

Second, and perhaps most strikingly, the gap was particularly pronounced in the arousal dimension. While valence judgments, the basic pleasantness of a word, remained relatively well aligned between learners and native speakers, the felt intensity of emotionally charged words was substantially attenuated in the instructional-setting sample. This points to emotional intensity as a critical ingredient in the acquisition and use of emotional language. Words learned through arousing, personally resonant experiences appear to acquire a physiological charge that classroom exposure struggles to reproduce. The finding dovetails with a growing body of evidence that the second language functions as a kind of emotional dampener, but it refines the picture by suggesting that what is dampened is chiefly arousal, the bodily component of emotion, rather than the evaluative component.

The methodological care underlying these conclusions deserves emphasis. The authors note that earlier comparable datasets suffered from subtle scale-direction confusions, and they deliberately standardized their procedures to avoid such artifacts. They also documented which words failed to gather sufficient valid observations, showing that rarer, later-acquired words were more likely to be reported as unknown by participants, an honest accounting of the vocabulary limits of an instructional population. All individual and averaged ratings, together with the study materials, have been released through the Open Science Framework, making the resource immediately usable by other researchers. Analyses were conducted in the open-source statistical package JASP, and the study, while not preregistered, was approved by the Human Research Ethics Committee of the University of Salamanca.

Why does this matter beyond the laboratory? Affective norms are the raw material of experimental psychology. Researchers designing studies of memory, attention, decision-making or reading rely on them to select stimuli, and using native speaker norms with learner populations risks systematic error, since a word the experimenter believes is arousing may register as flat to the participants. The new database gives scientists studying EFL populations a calibrated instrument for the first time at this scale, joining a growing international effort that has produced affective norms for Chinese, Dutch, German as a second language, Spanish and many other linguistic contexts. For educators, the implications are more practical still. Previous work by members of the same team found that EFL textbooks are strikingly poor in emotionally charged vocabulary, and the arousal gap documented here suggests a concrete mechanism by which that poverty limits learning. Teaching strategies that deliberately weave emotional intensity into vocabulary instruction, through personal narratives, vivid contexts and embodied activities, may help classroom learners close the gap that immersion closes naturally.

The study also resonates with a broader scientific debate about the relationship between language and emotion. Whether emotions are biologically basic categories, as classical theories propose, or constructed from linguistic and cultural ingredients, as psychological constructionist accounts argue, the new data show empirically that the linguistic environment in which a word is acquired shapes its emotional profile. Emotion semantics have been shown to exhibit both universal structure and cultural variation, and the Salamanca findings add an environmental variable to that equation: not just which language you speak, but where and how you learned it. For the world’s classroom-trained multilinguals, who vastly outnumber immersed bilinguals, the message of this research is both humbling and empowering. Their foreign-language words do carry less emotional voltage than their mother-tongue equivalents, and that asymmetry influences memory, moral judgment and everyday communication. But the asymmetry is measurable, tractable and, the authors suggest, addressable, provided that researchers and teachers design language experiences with the emotional intensity that the human lexicon seems to require.

Subject of Research: Affective norms for emotional word processing in college learners of English as a foreign language

Article Title: Emotion in English as a foreign language: Affective ratings for over 2,000 words by college learners

Article References: Bermúdez-Margaretto, B., Pérez-García, E., Fernandez, A., & Sánchez, M. J. (2026). Emotion in English as a foreign language: Affective ratings for over 2,000 words by college learners. Behavior Research Methods, 58(10), Article 291. https://doi.org/10.3758/s13428-026-03159-x

Image Credits: AI Generated

DOI: 10.3758/s13428-026-03159-x

Keywords: affective norms, valence, arousal, English as a foreign language, bilingualism, emotional word processing, second language acquisition, psycholinguistics, language learning, classroom learners, Emotion, English