Several decades ago, discussion of the prevalence of machine translation (MT) seemed almost unthinkable. For example, in 2001, Stanford professor M. Kay noted the insignificant progress of MT over the previous 40 years and, in his opinion, there were no prospects for the next 40 years (Youngblood, 2001).

Many foreign scientists held a similar view until the debut of Google Translate (GT). Today, decades after these pessimistic predictions, the smartphone market has exploded with translation apps that allow users to photograph text and translate it on the spot. As a result, translation has become an integral part of communication in the personal and academic lives of many people (Anderson, 2013).

It is believed that MT has a huge impact on English language learning. However, not everything is so clear-cut. An analysis of foreign theoretical literature on the use of machine translation in English language learning has demonstrated the existence of different points of view on this issue, both among students and teachers.

For example, in a survey conducted by A. Nino, 75% of students believe that MT is a useful language tool, and 81% said that MT helped improve their language skills (Niño, 2009). A study by J. Clifford et al. found that 88% of students used ML at least occasionally in their studies, of which 39% used it sometimes, 32% used it often, 17% used it rarely, and only 12% of respondents said they never used ML. It should be noted that the majority of MP users (81% of respondents) use GT as a tool to help them learn the language, and 33% of students consider GT to be a tool they could not live without.

It should be noted that some studies that argue in favor of using MP in language learning in an educational context start from the assumption that MP distracts from language learning, but its use by students is inevitable, and instead of focusing only on the possible abuses of this relatively new electronic tool, it would be better to study its potentially positive applications. Students need to learn to communicate in English, and in this sense, the use of machine translation is unnecessary, since text input and MT do not involve either communicative activity or language analysis. Nevertheless, the results of foreign studies indicate the widespread use of MT in the process of completing homework and written assignments (Williams, 2006).

In particular, an analysis of the purposes of MT use revealed that students most often use it to translate from English into their native language (96%). In addition, a high usage rate was also noted for the following grammatical and stylistic categories: vocabulary (91%), idiomatic expressions (36%), transitional words or conjunctions (31%), verbs of time (29%), and word order (20%). We would like to add that students used MT to work with individual words or small paragraphs to translate individual words (89%), short phrases of less than 5 words (62%), complete sentences (16%), and short paragraphs (7%). It is noteworthy that students indicated a tendency to sometimes or rarely use MP for the following tasks: reading, grammar homework, written homework, and essays.

It follows from the above that students use MP as dictionaries and to double-check what they have written in English.

We find the results of studies demonstrating students’ awareness of the limitations of MT accuracy to be interesting. Thus, it was found that 91% of users found errors when using MT. At the same time, 74.91% of students indicated that MT is only effective when translating individual words (Alhaisoni, Arabia, 2017).

When assessing the overall reliability of translations created with the help of MT, most respondents described them as very reliable (63.52%) and reliable (55.81%), significantly fewer as unreliable (28.13%) and very unreliable (6.35%) (Alhaisoni, Arabia, 2017).

All the reasons explaining how errors are detected in MT can be divided into three groups: knowledge of English; the result does not correspond to the original text; the result does not correspond to what was studied in class (see Table 1).

This point of view is also shared by a group of scientists (Kol, Skolnik, Spector-Kohen, 2018), who suggested in their work that MT can be useful for students if its results are critically evaluated and corrected.

The study found that when assessing written assignments in terms of grammatical accuracy, there were no significant differences (93.2 points without MP and 93.4 with MP). When comparing the average number of words in the written assignment, it was found that students wrote significantly more words when using MP (t=-2.04947. p = .023348 at p <.05). In addition, it was found that the average score for the first written assignment (without MP) was 8.6, and for the second assignment (with MP) – 10.3. When comparing the two averages, it was found that when using MP, the sentences were more common (t = -2.71851. p=.004745 at p < .05) (Kol, Schcolnik, Spector-Cohen, 2018).

During the experiment, the words in the texts written by the students were divided into three groups: K1, the first 1,000 most frequently used words in English; K2, the next 1,000 most frequently used words; and AWL, the academic word list (see Table 2).

Table 2 shows that when using MP in the writing assignment, students’ vocabulary improved, i.e., K1 words decreased, while K2 and AWL words increased. Tests were conducted for comparison. Thus, between the first and second writing assignments, significant differences were found for all three groups of words. For K1, t = 8.60713. p < .00001 (at p < .05), which means that students used fewer basic words when writing with MP. For K2, t = -4.78664. p < .00001 (at p < .05), demonstrating that students use more words from the second group when writing with MP. For AWL, t-2.18186. p = .017133 (at p < .05), showing that students used more academic words when writing with MP.

From the above, it follows that the number of lexical units used by students when writing written assignments with MP, K2, and AWL increased significantly, while the number of KI words decreased.

The data show that students wrote significantly more words when using MP in the process of completing written assignments. This does not indicate that the words have become part of their productive vocabulary, but it does not rule out the possibility that respondents will expand it over time, since it is known that the quality of academic writing depends heavily on the use of academic vocabulary (Nation, 2001).

Equally interesting is the position of teachers regarding the use of MP in the process of learning English in higher education institutions.

According to A. Nino, 23% of teachers used MP in their classes. At the same time, 30% of teachers who did not use MP said that they were ready to do so in the near future. It is noteworthy that in his work, A. Nino studied not only the characteristics of teachers’ perceptions of the use of MT among their students, but also their understanding of teachers’ attitudes towards new translation tools.

To determine teachers’ attitudes toward this issue, respondents were asked the question, “Is it cheating for students to use MT?” Forty-two percent answered “yes,” 37% chose another category, and 21% of respondents answered “no.” Some teachers also mentioned that they consider it cheating when students present work as their own after using MP. In addition, 77% of teachers chose the option “disapprove” or “strongly disapprove” when asked about students’ use of MP; 23% chose the option “disapprove, but do not condemn”. It is worth noting that no teacher chose the answer “approve” or “strongly approve” when answering the above question.

In a 2012 survey of foreign language teachers at a regional Swedish university, 66% of respondents said they would prefer their students not to use MP when completing written assignments. However, all teachers agreed that if students used MP, they would need good language skills to edit the results.

In 2013, a research group at Duke University studied teachers’ attitudes toward the use of MT by students in the process of learning English. It was determined that 77% of respondents did not approve of the use of machine translation by students. 84% of teachers teaching beginners believed that MT was not a useful tool. However, 54% of teachers working with students who had advanced English skills believed that MT was useful for language acquisition. At the same time, 42% of teachers considered the use of MT in written assignments to be cheating (Clifford, Merschel, Munne, Reisinger, 2013).

We would like to highlight the attitude of teachers towards the personal use of MP for academic or non-academic purposes. Analysis of the data clearly shows that only 5% of teachers used MP for academic purposes and 7% for non-academic purposes (see Table 3).

When asked how often teachers talk to their students about using MP for academic purposes, 72% responded that they do so at least once per semester, 19% responded that they do so depending on the situation, and 9% responded that they never do so. Some of the respondents explained that they had never thought it was a relevant topic for discussion (Niño, 2009).

Of particular interest are the data on teachers’ perceptions of the usefulness of ML in the language learning process. Thus, 7% of teachers responded that it is “useful,” 33% “useless,” and 60% chose the option “depends on the situation.” An analysis of the results obtained in the study on whether MT is useful in the language learning process depending on the level of proficiency shows that teachers consider machine translation tools to be useless or unsuitable for language learning (see Table 4).

However, it should be noted that for students with a more advanced level of language proficiency, teachers consider MT to be a more useful tool (see Table 4). Thus, the survey results confirm the previously put forward hypotheses that MT can only be used effectively when students have a good command of English and are able to spot mistakes and learn from them (Niño, 2009).

The data presented demonstrate the ambiguity in teachers’ perceptions of the role of MT in the process of learning English. On the one hand, teachers view MT as a form of cheating or as an unsuitable tool for language learning; some teachers fear that MT will contribute to the elimination of foreign language learning programs. On the other hand, teachers suggest greater integration of MT in the process of learning English and demand recognition of the existence of such translation tools (Niño, 2009).

1. Baker suggests that “the use of online translators can also be seen as a form of language socialization” (Baker, 2013); M. Pena argues that MT helps beginner students work with a larger number of texts and interact more with the English language; M. Case also points to the usefulness of MT for presenting authentic language materials to beginner and intermediate students; L. Williams, in turn, believes that the use of MT can “make students think of language as a tool for communication rather than a set of words or phrases”.

Thus, the growing interest in the role that MT plays in foreign language education is a natural response to changes in a society that is becoming increasingly globalized and where languages play an important role. As a consequence of growing multiculturalism, the use of machine translation by students is a topical issue discussed in academic circles.

A theoretical analysis of foreign sources on this issue shows that students regularly use MT to find words, translate assignments, and double-check their work. Students perceive MT as useful in language learning, especially in terms of expanding their vocabulary. Teachers, on the other hand, are skeptical about the positive impact of MT on English language learning. They believe that when using MT as a dictionary to look up individual words, students miss out on learning about the nuances and translation alternatives offered by more traditional dictionaries. However, some teachers still predict the integration of MP into the process of learning a foreign language and, as a result, the emergence of a demand for the recognition of such translation tools.