Computers and Education: Artificial Intelligence 5 (2023) 100156

Contents lists available at ScienceDirect

Computers and Education: Artificial Intelligence

journal homepage: www.sciencedirect.com/journal/computers-and-education-artificial-intelligence

Exploring generative artificial intelligence preparedness among university

language instructors: A case study
Lucas Kohnke a, *, Benjamin Luke Moorhouse b, Di Zou c
a
The Education University of Hong Kong, Department of English Language Education, Room: B4-1/F-07, 10 Lo Ping Road, Tai Po, Hong Kong
b
Hong Kong Baptist University, Department of Education Studies, AAB 830 Academic & Administration Building, Baptist University Road Campus, Hong Kong
c
Lingnan University, Hong Kong

A R T I C L E I N F O A B S T R A C T

Keywords: The integration of generative artificial intelligence (AI) in English language teaching presents opportunities and
Artificial intelligence challenges for instructors. This study explores the attitudes of higher education English language instructors
AI towards generative AI tools, their intentions to use them and the institutional support and professional devel­
Generative AI
opment necessary to teach and learn with them. As the field continues to evolve rapidly, it is essential to
University language instructors
Higher education
comprehend the readiness of front-line language instructors. This qualitative interpretive study seeks to identify
English the digital competencies and pedagogical knowledge required to implement generative AI in education and
provide guidance for the design of professional development programmes that address the challenges and
concerns associated with adopting AI. Drawing on semi-structured interviews with twelve instructors at a higher
education institution in Hong Kong, the findings reveal the significance of familiarity and confidence with using
AI-driven teaching tools, the challenges and concerns language instructors face and the need for tailored support
and professional development. The study offers ten practical implications to cultivate language instructors’
digital competencies, pedagogical knowledge and positive attitudes towards integrating AI to enhance their
students’ learning experiences.

1. Introduction Recent studies have investigated how traditional AI can be used to

enhance education by optimising face-to-face, blended and online in­
While artificial intelligence (AI) has enhanced living, learning and struction (Kexin et al., 2020), reducing workload by automating tasks
working environments and become a part of daily life (in smart home (Weng & Chiu, 2023) and enhancing the data analysis process (Euro­
appliances, chatbots and more), the new wave of generative AI tools pean Commission, 2022). AI-powered tools can also assist instructors in
based on Large Language Models (e.g. ChatGPT, DALL-E2) is set to identifying effective pedagogies based on student learning data, gener­
revolutionise many domains. In education, scholars and practitioners ating assessments and issuing grades and feedback automatically
have recently begun to seriously discuss their implications for teaching, (Chaudhry & Kazim, 2021). In addition, research has suggested that AI
learning and assessment (Kohnke et al., 2023). Despite increases in ed­ can promote personalised and adaptive learning (Chiu et al., 2023;
ucators’ digital competencies and the proliferation of technology use Kabudi et al., 2021), accelerate knowledge acquisition and motivate
triggered by the COVID-19 pandemic (Howard et al., 2022; Moorhouse, students (Chen, Xie, Zou, & Hwang, 2020a, Chen et al., 2020; Hwang
2023; Sartika et al., 2021; Whalley et al., 2021), language instructors et al., 2020). However, teachers may not have the technical skills to use
will likely need AI-specific digital competencies to implement these educational applications driven by generative AI (e.g. Ally, 2019; Seo
tools appropriately (Kohnke et al., 2023). Generative AI tools represent a et al., 2021). The challenges of applying traditional AI in education
dramatic advancement from previous AI tools. It leverages deep learning include its limitations, misconceptions among teachers and their general
models to generate human-like content, including audio, code, images, lack of technical expertise (Akgun & Greenhow, 2021; Sijing & Lan,
text, simulations, 3D objects and videos, including new and unexpected 2018), all of which will likely be exacerbated by generative AI.
outputs in response to varied and complex prompts (e.g. languages, Teachers may benefit from capitalising on generative AI tools to
instructions, questions; Lim et al., 2023). design and support learning experiences. However, their influence on

* Corresponding author.
E-mail addresses: [email protected], [email protected] (L. Kohnke), [email protected] (B.L. Moorhouse), [email protected] (D. Zou).

https://doi.org/10.1016/j.caeai.2023.100156
Received 15 June 2023; Received in revised form 30 June 2023; Accepted 12 July 2023
Available online 13 July 2023
2666-920X/© 2023 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
L. Kohnke et al. Computers and Education: Artificial Intelligence 5 (2023) 100156

education is still uncertain (Holmes et al., 2022s required to understand technologies and equip them with the digital competencies needed to
whether and how these emerging technologies will benefit teachers. enrich students’ learning experiences.
Previous studies have focused on fostering students’ AI competencies (e. Teachers have multifaceted perceptions and concerns relating to
g. Kong et al., 2021; Xia et al., 2022). However, it is equally important to adopting AIED. Some fear being replaced, assuming that AI will weaken
consider how the competencies of language teachers change in an AI their professional positions (Holmes & Tuomi, 2022). This can
context and how they adopt AI-driven tools and approaches. This contribute to a reluctance to embrace AI technologies, as teachers may
qualitative study explores the attitudes of twelve higher education En­ fail to perceive that AI can complement rather than replace them. In an
glish language teachers toward generative AI tools, their intentions to era marked by increasing awareness of the ethical implications of
use them and the institutional support and professional development technology, teachers may also be hesitant to adopt AI-driven tools
needed to facilitate the transition to AI-assisted teaching and learning. without a clear understanding of how student data will be protected and
As this is a rapidly developing field, it is essential to gain a qualitative algorithmic biases minimised. Furthermore, in many settings, a lack of
understanding of the readiness of front-line teachers. infrastructure, funding and support hinders the widespread adoption of
AI (Popenici & Kerr, 2017; Sharma et al., 2022). Lastly, teachers may not
2. Literature review possess the technical or pedagogical knowledge and expertise to inte­
grate AI-driven tools into the classroom, leading to apprehension and
In 1956, AI was conceptualised as ‘the science and engineering of reluctance.
creating intelligent machines’ (McCarthy, 2007, p. 2). Newell and Si­ For any or all of these reasons, teachers may feel ill-equipped to
mon’s (1956, p. 868) ‘thinking machine’, developed in the same year, leverage AIED tools to enhance their instructional practices and improve
was the first computer program to simulate human intelligence by student learning. Some universities in Hong Kong have banned ChatGPT
addressing complex problems. Since then, researchers and educators for students (Yau & Chan, 2023) and few have offered teachers sys­
have contributed to the field significantly by employing AI in Education tematic training on the technical knowledge they need to integrate AIED
(AIED) – for example, by using AI-based technologies, applications and tools effectively or an understanding of the pedagogical uses of AI. This
programs as intelligent tutors, tools/partners and policy advisors to may result in the suboptimal application of AIED.
facilitate teaching, learning and decision-making (Hwang et al., 2020). To overcome these roadblocks, teachers must be equipped with the
Given the potential of AI to enhance education, researchers, policy­ digital competencies needed to implement AI in a pedagogically sound
makers and practitioners have begun to focus on AIED in recent decades manner (e.g. developing learning tasks that take advantage of AI; using
(e.g. Williamson & Eynon, 2020). They have developed AI-driven ChatGPT-4 to personalise and plan lessons; Hrastinksi et al., 2019). As
learning tools, including intelligent tutoring and recommendation sys­ Hockly (2023) contends, it is also critical to emphasise the ethical use of
tems. These innovations allow AI to evaluate and enhance student AI technologies and develop strategies to mitigate the potential draw­
learning experiences by providing personalised guidance (Hwang et al., backs. We would further argue that the competencies needed to use
2020; Williamson & Eynon, 2020). For example, it can provide students generative AI tools effectively go beyond those previously required to
with learning materials or paths tailored to their needs (Christudas et al., integrate technology into teaching (Kohnke, Moorhouse, & Zou, 2023).
2018), diagnose their strengths, weaknesses and knowledge gaps (Liu Several approaches can be adopted to support teachers in developing
et al., 2017) and identify those who are at risk of failure (Luckin et al., competencies related to generative AI. First, continuing professional
2022). development opportunities can help teachers gain technical skills and
Recent studies of AIED have explored technologies and tools (e.g. knowledge (Ng et al., 2023). Second, fostering a positive attitude to­
Zhai et al., 2021), their geographical distribution, the subject areas in wards AI and emphasising its role as a supportive tool rather than a
which they are used and patterns in the textual data (Bozkurt et al., replacement for teachers can alleviate their concerns and encourage
2021). Others have analyzed the use of AI in specific disciplines such as adoption (Qian et al., 2021). Teachers should be assured that AI will
language, mathematics or medicine (Karaca et al., 2021) or assessments complement their expertise and only be used to automate routine tasks
(González-Calatayud et al., 2021). Some have investigated particular (Gao, 2021). Third, establishing clear guidelines and policies regarding
technologies or applications, such as assistive robots or adaptive the privacy of data and ethical usage can address common concerns and
learning (Nigam et al., 2021; Papadopoulos et al., 2020). Chiu et al. encourage responsible implementation (OECD, 2022). Finally,
(2023) recently examined the opportunities and challenges associated providing teachers with funding to access resources and infrastructure
with four dimensions of AIED: (1) employing AI for students’ learning (e. can also facilitate its adoption (Zhang & Aslan, 2021). This may require
g. facilitating human–machine conversations, analysing student work, securing collaboration from governments and private organisations.
offering feedback, enhancing adaptability and interactivity); (2) utilis­ In summary, integrating AI technologies into education holds the
ing AI for teaching (e.g. developing adaptive strategies, improving potential to revolutionise the teaching and learning experience. While
teachers’ capabilities, supporting professional development); (3) recent publications have explored various aspects of AIED, teachers lack
implementing AI for assessment (e.g. automating grading, predicting the digital competencies to implement it in the classroom. Teachers’
student performance; (4) integrating AI into educational willingness to adopt AIED tools is influenced by their perceptions,
decision-making (e.g. providing evidence-based insights and conve­ concerns and access to resources, as well as other factors. Therefore, it is
nient, personalised services). However, few studies have focused on crucial to equip them with the digital competencies necessary to harness
identifying the digital capabilities instructors need to navigate AI-driven the affordances of AI technologies. This explorative study will address
learning environments. More importantly, the AI tools and chatbots the following questions:
deployed in these studies were not generative AI, so the new technology
remains unexplored. 1. What digital competencies and pedagogical knowledge do teachers
According to Ng et al. (2023), AI technologies offer educators novel need to implement AI in teaching?
avenues to enhance their instructional methods, including personalised 2. How can professional development programmes address teachers’
support, seamless communication and learning analytics. By employing challenges and concerns related to adopting AI technologies?
AI-based tools, teachers can not only improve their efficacy (Healy &
Blade, 2020) but also foster self-regulation among their students (Seo This study will contribute to the limited literature in this area,
et al., 2021) and facilitate meaningful communication and interaction providing teachers and policymakers with strategies to promote and
(Torda, 2020). Despite these benefits, many educators have not yet implement generative AI in education.
adopted AI-driven technologies. Consequently, it is imperative to un­
derstand the obstacles they might encounter when implementing such

2
L. Kohnke et al. Computers and Education: Artificial Intelligence 5 (2023) 100156

3. Methodology • How do you think your current digital competencies and pedagogical
knowledge have influenced your ability to use AI in teaching?
This study was situated within the qualitative interpretive paradigm • What are the most important topics or skills to include in a profes­
(e.g., Denzin & Lincoln, 2011), as it aimed to elucidate the diverse sional development programme focused on AI in education?
perspectives of English language instructors. The research was con­ • What types of training or support have you found effective in helping
ducted through semi-structured interviews at an English-medium uni­ you develop the skills needed to implement AI in teaching?
versity in Hong Kong during the Spring of 2023. The goal was to elicit
in-depth commentary on the lived experiences and perceptions of the The interviews were conducted in English and lasted 28–42 min.
participants (Creswell, 2008). They were audio-recorded and subjected to manual iterative thematic
analysis (Braun & Clarke, 2006) to yield a comprehensive, detailed and
rich understanding of the data. The recordings were first transcribed and
3.1. Research context and participants
shared with participants in a preliminary member check (Merriam,
Tisdell, & E, 2016). Then, the three authors carefully reviewed and
This study included twelve instructors from the English language
re-examined the transcripts to familiarise themselves with the data.
centre at a publicly-funded university in Hong Kong as participants. This
Subsequently, they independently coded the transcripts, generating
university is one of eight institutions offering
initial codes that were shared, discussed and refined using Google Docs.
undergraduate and postgraduate programmes across a wide range of
They reached a consensus on the themes and sub-themes. A code–recode
disciplines, all taught in English. The English language centre plays a
strategy was implemented to enhance the consistency and dependability
crucial role in supporting students’ transition to university life and the
of the findings (Anney, 2014). All three authors revisited the data and
English-medium environment.
re-coded it two weeks after the initial coding. The integrity of the
The participants had similar educational backgrounds: either a
findings was confirmed by the near-identical outcomes. A second
master’s or doctoral degree in English language teaching. There were six
member check was performed to further improve the reliability of the
male and six female instructors. Their ages ranged from 36 to 57, and
data. Participants were provided with a summary of the study’s findings,
they had between 10 and 32 years of teaching experience. All partici­
including the identified themes and representative quotations (Merriam
pants signed consent forms and were assigned pseudonyms. Although
et al., 2016). None of the language instructors suggested or requested
some had experience with AI tools in their personal lives, none had
any amendments in either of the two member checks.
previously integrated AIED into their classes. See Table 1 for a profile of
participants.
4. Results

3.2. Data collection and analysis We have organised the results around the four main themes that
emerged during the analysis. Themes 1 and 2 provide insight into RQ1,
This research drew on an interpretive paradigm to gain insight into and Themes 3 and 4 address RQ2 (see Table 2). In this section, the
instructors’ perceptions of and experiences with AIED. A qualitative language instructors’ interview responses are presented verbatim.
research design was employed to address the two research questions,
with individual semi-structured interviews enabling in-depth explora­ 4.1. Theme 1: unconscious familiarity with AI and its impact on attitudes
tion of the subject (Cohen et al., 2011). Convenience sampling was
adopted in this study, with the first author inviting instructors at one of Overall, the participants were familiar with AI tools and some of
the English language centers Hong Kong to participate. The following their functions. They interact with AI tools daily and mentioned appli­
definition was used: ‘Generative artificial intelligence (AI) describes cations such as Siri, Alexa, Google Assistant, Netflix and Fitbit. The in­
algorithms (such as ChatGPT) that can be used to create new content, terviews revealed that they use the tools to streamline tasks and
including audio, code, images, text, simulations and videos’ (McKinsey personalise their experiences. Oliver said that Siri helps him keep track
& Company, 2023). The interview guide comprised the following
questions: Table 2
Summary of themes and relationship with the research questions.
• What digital competencies do you think English language instructors RQ1: Theme 1: Unconscious Theme 2: The
need to use AI in their teaching practice effectively? What digital familiarity with AI and its importance of
• Can you identify any key pedagogical considerations for integrating competencies and impact on attitudes familiarity and
AI into the classroom? pedagogical knowledge Highlights language confidence to
do instructors need to instructors’ past and effectively integrate
• What are the main challenges you face when adopting AI technolo­
implement AI in present interactions with generative AI into
gies in your teaching? teaching? AI technology and how teaching practices.
this familiarity shapes Emphasises the
Table 1 their attitudes towards necessity that language
integrating AI in the instructors become
Profiles of the participants.
classroom. familiar and confident
Participant Gender Age Education Teaching Experience with generative AI tools
(Pseudonym) Experience with AIED before integrating them
into teaching.
Noah M 41 MA 14 No
Oliver M 39 EdD 12 No RQ2: Theme 3: Tailored Theme 4: Adopting AI
Henry M 53 MA 27 No How can professional support for developing in the subject or
Amelia F 36 MA 8 No development instructors’ AI teaching discipline
Olivia F 46 EdD 19 No programmes address the competencies Considers the
Evelyn F 41 MA 15 No challenges and concerns Highlights the help challenges instructors
Sophia F 52 PhD 25 No of adopting AI instructors need to use AI face, including the need
Will M 37 MA 11 No technologies? in teaching skillfully (e.g. for specific knowledge
Theodore M 36 PhD 10 No hands-on workshops, self- about using AI to teach
Mia F 48 MA 22 No paced learning language, ethical
James M 57 MA 32 No experiences, personalised concerns, plagiarism
Isabella F 43 MA 16 No coaching). and digital citizenship.

3
L. Kohnke et al. Computers and Education: Artificial Intelligence 5 (2023) 100156

of his daily schedule by reminding him about important work and per­ ‘develop their communicative skills’ and ‘learn independently outside
sonal tasks. Sophia and Mia mentioned using Siri and Google Maps to the classroom’. However, they conceded that this will be difficult if
check traffic updates and weather forecasts. These tools, however, have students rely on AI tools excessively. The interviewees were generally
become so familiar that the language instructors do not always equate positive about the potential of generative AI tools in education but un­
them with AI. For example, Amelia stated, ‘I use[my] Fitbit daily and did certain about how to apply them. This suggests that professional
not consider it AI but another technology. It was not until recently, with development providers should include strategies for incorporating such
the launch of ChatGPT, that I also started to consider this AI.’ tools to promote deeper learning and enhance students’ critical thinking
Although AI has long been embedded in the participants’ daily lives skills.
through such interactive tools, the development of ChatGPT and similar
programs made them more conscious of AI and its implications. For 4.3. Theme 3: Tailored support for developing instructors’ AI teaching
example, James said that after the ‘launch of ChatGPT, everyone began competencies
discussing AI suddenly’. Similarly, Noah suddenly realised he had been
“utilising various sorts of AI technology for quite some time”. Will noted, The third theme concerned the support participants need to become
‘Before this AI hype, I wasn’t even giving it a second thought, but now I proficient at teaching with AI. Most participants agreed that their uni­
realise I have been using it for years.’ Thus, the participants were versity has not done enough in recent months to help them keep up with
familiar with AI-enabled tools, such as virtual assistants and fitness the accelerating development of generative AI tools; it has solely relied
applications, and used them frequently, but they did not realise they on communicating ambiguous guidelines by email and has not offered
were using AI until the release of ChatGPT. Understanding language opportunities for dialogue. The language instructors suggested that a
instructors’ experiences with AI can lead to professional development hands-on workshop with opportunities to explore and use AI tools in a
programmes that build on their knowledge and familiarity to make them supportive environment would be preferable. Mia said, ‘There is too
amenable to adopting generative AI in their classrooms. much discussion about the positive and negative aspects of ChatGPT-4 in
the hallways and in emails from the university. Still, we don’t receive
4.2. Theme 2: The importance of familiarity and confidence to effectively training on effectively using it in workshops or seminars.’ Several par­
integrate generative AI into teaching practice ticipants agreed with Mia on the need for informal workshops that allow
them to try out different prompts and discuss how they can complement
The second recurring theme was that the participants need to their existing materials and assessments. Will suggested that the
become familiar with generative AI tools before they can consider using
university arrange self-paced learning experiences where they work
them in their teaching practice. Isabella explained, ‘Before I can deter­
through different scenarios to get a better understanding. And they
mine whether it is appropriate for my students, I need first to try and
should make it easy for us to access articles, journals, and reviews on
figure out how it works’. Henry added, ‘My job is to teach my students
the website.
how to write academic essays, so I need to know how ChatGPT-4 can
supplement rather than replace my teaching.’ When asked to elaborate, Amelia proposed ‘informal sharing sessions’ to provide a platform for
he shared that he is ‘uncertain how to design learning activities that ‘everyone to be honest about their concerns’. The participants agreed
would help learners to think and develop their writing skills critically they need more information to understand the benefits and limitations
and not simply copy and paste’. Two participants, Olivia and Theodore, of AI for the classroom. In Sophia’s words, ‘We need to learn more about
have developed preliminary ideas of how students could use ChatGPT-4; how it can enhance our instruction and not replace it.’
as Olivia explained: ‘Before, I instructed my students to start with a Another suggestion was that language centers and universities pro­
Google search to gather background information; next semester, vide personalised coaching and mentoring led by language instructors
perhaps, ChatGPT-4 would be a better way for them to generate ideas.’ who are proficient in AI tools. Henry commented, ‘During the COVID-19
However, both agreed that they need to gain more confidence with pandemic, we teamed up with colleagues to ask for guidance on using
ChatGPT-4 before deploying it in their teaching. technology. I think they should promote this now also, as it would help
Several other participants mentioned similar pedagogical consider­ some of us to catch up.’ When asked to explain how to accomplish this,
ations. Evelyn said, ‘Everything is new for both students and us as ed­ Olivia suggested small, private online courses focusing on different as­
ucators; we need to learn how to use AI effectively to accomplish the pects of AI:
desired learning outcomes.’ Most participants agreed they are not pre­
One could be ethical concerns about using AI-generative tools, while
pared to utilise generative AI tools to improve teaching and learning.
another could be how to use them to teach and assess speaking and
Henry, who considered himself highly proficient at using technology,
writing. Personally, I think this would be an excellent method to help
described generative AI as ‘a whole new world’ that requires a new skill
us become familiar with and learn more in-depth in our own time
set and makes it ‘hard to keep up’. Most participants said that their
anywhere.
current level of digital competency influences the degree to which they
can use AI. For example, Noah shared that his This theme highlights the need for various formats and modes of
professional development that cater to language instructors’ individual
limited experience using AI tools will make it challenging to identify
needs and learning styles and enable them to develop the skills and
the best ways to incorporate them into my teaching practice in the
knowledge needed to integrate generative AI effectively.
coming semester. Using ChatGPT for lesson planning and writing
quizzes is easy. Still, I think my students would also like to use it in
4.4. Theme 4: Adopting AI in the subject or discipline
completing assignments.
When asked to describe the specific technological and pedagogical The fourth theme related to the specific knowledge English language
resources they would need to teach in an AI environment, the partici­ instructors need to incorporate AI tools into their teaching, as some tools
pants responded that they want to become more familiar with using have subject-specific uses or challenges. Will said he wants to use
chatbots to help their students practise writing. Specifically, they want generative AI tools ‘to provide individualised feedback to personalise
to know ‘what prompts to use to help best students develop their learning’ and offer ‘targeted support for weaker students for specific
writing’ (Olivia). grammar points such as incorrect verb tense usage and vocabulary
Other participants said AI could help their students engage with their building’. He elaborated: ‘Having more information about chatbots and
work and learn more effectively or anticipated that it could provide intelligent tutoring systems would be helpful’ for such activities. Simi­
automatic feedback. Oliver and Isabella noted that their students need to larly, James claimed that AI tools could help students improve their

4
L. Kohnke et al. Computers and Education: Artificial Intelligence 5 (2023) 100156

pronunciation and communication skills, allowing them to ‘practise teachers’ AI teaching competencies, such as hands-on workshops, self-
speaking skills using authentic language’. However, the language in­ paced learning experiences and personalised or individualised coach­
structors felt that the few tools applicable to teaching English with ing. This is consistent with the recommendations in the literature
which they are familiar (e.g. Loora.ai and SmallTalk2.me) are too (Hrastinski et al., 2019) and suggests the necessity of securing support
expensive. This suggests that professional development programmes from governments and private organisations to provide the necessary
should include content specific to language instructors’ subject areas. funding and resources (Ng et al., 2022).
The participants also identified some issues that may arise from using
AI in language classes. They discussed the importance of addressing 6. Conclusion and implications
plagiarism and the implications of using AI-generated text for academic
assignments. Olivia stated she is ‘worried that students won’t do the In conclusion, this study contributes to the understanding of what
work independently’ and Sophia added, ‘It’s so simple for students to try language instructors need (in terms of digital competency, pedagogical
to cheat.’ In response to the concern that students would turn to knowledge and professional development) to integrate generative AI in
ChatGPT-4 to complete their written assignments, Will suggested that education. The findings offer practical implications for the design of
language instructors could help students learn how to ’evaluate online professional development programmes that address teachers’ challenges
information and use it responsibly’. They agreed that their job requires and concerns and thus help them use such tools effectively. The
them to teach language skills and foster digital citizenship among their following recommendations are designed to help educators, policy­
students. makers and stakeholders create environments in which AI technologies
In the interviews, participants expressed curiosity about how AI are effectively harnessed to enhance the teaching and learning
could be implemented. Most agreed that it would be impossible to ban experience.
ChatGPT-4. However, Evelyn mentioned a concern about the additional This section offers nine practical implications and recommendations
workload and time required ‘to figure out how to apply it in the best way for developing English language instructors’ digital competencies and
for students’ language learning’ (Evelyn). In a similar vein, Amelia pedagogical knowledge so they can implement AI in teaching. They are
discussed the need to change her teaching approach to ‘accommodate AI arranged in terms of three main themes.
tools in her classroom’. These points highlight the importance of
designing professional development activities that address these issues 6.1. Theme 1: Education and training on using AI in language teaching
and guide language instructors to navigate them within their particular
subject or discipline. (1) Emphasise the role of AI in language education: Highlight its
potential to supplement traditional teaching methods by offering
5. Discussion individualised instruction, feedback and assessment. Some
possible strategies for implementation involve inviting expert
The findings provide valuable insight into the digital competencies speakers or organising webinars in which individuals share in­
and pedagogical knowledge language instructors need to effectively sights about AI’s benefits and practical applications in language
implement AI tools in the classroom. They also illustrate the obstacles teaching. However, English language instructors may resist new
teachers face and the support they need. As stated in the literature re­ technologies due to their perceived complexity or lack of rele­
view, teachers’ attitudes towards AI integration in the classroom are vance to their current practice. To mitigate this, it is crucial to
substantially affected by their familiarity with the technology. This emphasise the tangible benefits of AI by showcasing real-world
study confirms that language instructors’ familiarity with AI is often success stories of its application in the classroom.
unconscious because it is embedded in so many common digital tech­ (2) Offer hands-on training about AI for language teaching:
nologies. The launch of ChatGPT and the resulting media attention ap­ Conduct workshops and practical sessions in which English lan­
pears to have increased language instructors’ awareness of AI and made guage instructors interact with AI tools and applications related
them think about it more productively and critically. to language instruction, such as automated essay graders, adap­
Despite this, the skills needed to use AI tools for teaching seem un­ tive learning platforms and conversation practice tools. These
derdeveloped in the participants, who explicitly stated that they lack the sessions should be led by AI specialists who can provide clear,
confidence to integrate them and teach students how to use them pro­ comprehensive training. One potential challenge is that language
ductively and responsibly. In line with the relevant literature (e.g. instructors may feel overwhelmed by the technical aspects of AI.
Christudas et al., 2018; Liu et al., 2017), the findings of this study sug­ Tailoring the training to their level of technical proficiency and
gest that language instructors must be familiar and confident with using providing support materials and follow-up sessions are, therefore,
generative AI tools to apply them effectively. They also need to develop essential.
digital skills and a pedagogical understanding of such tools. Addressing (3) Create self-paced learning opportunities for language
these needs can rectify the shortcomings identified by Chiu et al. (2023) teachers: Design online courses and self-study modules about AI
and Ng et al. (2023) The results also highlight challenges and concerns tailored to the needs of English language instructors that give
language instructors face when adopting AI technologies, such as the them the flexibility to learn at their own pace. These resources
fear of being replaced and ethical considerations (Du & Gao, 2023). should be easy to access and have user-friendly interfaces.
According to Hockly (2023), the ethical use of AI and developing stra­ However, teachers may lack the motivation to complete self-
tegies to mitigate its potential drawbacks are essential components of paced courses due to their busy schedules. Encouraging regular
professional development programmes. If their concerns are addressed, use by integrating these resources into existing professional
language instructors can be encouraged to view AI as a useful supple­ development programmes and providing recognition or in­
ment to their instruction rather than a threat (e.g. Ratten, 2020). centives for course completion may help increase engagement.
Research on the effect of generative AI tools on education is still in its (4) Integrate AI into teacher education programmes: Include AI-
early stages and language instructors are still ‘finding their footing’. related competencies and skills in the curricula of pre-service
Accordingly, there is a pressing need for professional development teacher education programmes. This can be achieved by collab­
programmes specifically designed to cultivate positive attitudes and orating with the curriculum designers. However, finding educa­
facilitate the transition to AI-driven English language teaching practices. tors who are competent in both language teaching and AI may be
These programmes should recognise and build upon teachers’ previous challenging. Potential solutions include providing training for
AI experiences, as described in this study. educators to build their AI knowledge or partnering with AI
Finally, this study highlights the need for tailored support to develop specialists who can co-teach or guest lecture.

5
L. Kohnke et al. Computers and Education: Artificial Intelligence 5 (2023) 100156

(5) Continuously evaluate and update professional development the interests and priorities of diverse stakeholders could be a
programmes: Implement systematic evaluation mechanisms, challenge, as each party may have a unique perspective on the
such as surveys, interviews and classroom observations, to assess role of AI in language education. To address this, maintaining
whether professional development programmes enhance the AI- open lines of communication, clarifying roles and responsibilities
related competencies of English language instructors. This and developing shared objectives could promote a collaborative
continuous feedback can suggest areas of improvement and drive approach.
refinement. However, creating an effective evaluation system
may be challenging due to the diverse learning experiences and 7. Limitations of the present study
outcomes of English language instructors. To overcome this, a
combination of qualitative and quantitative evaluation methods The limitations of this study include its relatively small sample size,
could be used to provide a holistic understanding. Furthermore, which may restrict the generalizability of the findings. Despite this, the
professional development programmes should be updated regu­ sample size is sufficiently large to generate
larly to align with the evolution of AI technologies and the a ‘new and richly textured understanding’ (Sandelowsky, 1996, p.
changing demands of language education. To do so, it would be 183) of the phenomenon under investigation. Within the interpretive
beneficial to establish a dedicated team or task force that stays paradigm, the focus is on relevance rather than rigor at an axiological
abreast of these changes and ensures that the content of profes­ level. As such, the applicability of the results to a reader’s professional or
sional development programmes is current and applicable. personal context must be assessed on an individual basis (Merriam et al.,
2016). Regarding the representativeness of the language instructors, it
6.2. Theme 2: Collaborative learning and peer support should be noted that they represent the general educational background,
age and gender of language instructors at higher education institutions
(1) Encourage peer support and collaboration among language in Hong Kong. This helps to capture a broad range of experiences and
instructors: Foster a collaborative environment where English perspectives. While the sample size may limit the global generalizability
language instructors can share their experiences, learn from each of the results, the study provides valuable insights into the phenomenon
other’s successes and challenges and support each other to within the Hong Kong context.
incorporate AI into their classrooms. This could be achieved by However, this study provides a strong foundation for future research.
setting up learning communities or peer-coaching initiatives By exploring the perspectives of English language instructors in higher
where English language instructors can exchange ideas, share education regarding the digital competencies, pedagogical knowledge
best practices and solve problems collaboratively. However, time and professional development required to integrate generative AI in
constraints and competing responsibilities may deter language teaching, the findings contribute to the design of effective professional
instructors from participating active. To address this, it might be development initiatives. Such programmes can resolve instructors’
helpful to create flexible structures of participation, such as on­ challenges and concerns, thereby facilitating the successful integration
line forums or asynchronous discussion boards, that allow En­ of generative AI tools into English language classrooms. Future research
glish language instructors to engage at their convenience. could use larger samples and incorporate diverse methodologies, (e.g.
(2) Provide personalised coaching: Implement a mentor system quantitative, mixed-method and observational approaches), to broaden
where English language instructors receive guidance in applying the scope.
AI technologies to their unique teaching contexts and learning
objectives. Experienced educators or AI specialists could serve as Data availability
mentors, providing one-on-one support to English language in­
structors as they navigate the process of integrating AI into their The dataset generated and/or analyzed during the current study are
instruction. The main challenge here might be finding a sufficient not publicly available due to privacy policy.
number of mentors with the necessary expertise and free time.
Therefore, building partnerships with technology companies or
Code availability
universities could be a beneficial strategy to recruit mentors or
provide training for potential mentors within the teaching
All code included in this study is available from the first author upon
community.
reasonable request.

6.3. Theme 3: Policy, ethical guidelines and stakeholder engagement

Funding
(1) Address ethical concerns: Develop guidelines and policies
about data privacy, algorithmic biases and the responsible use of Not applicable.
AI. These policies should be informed by input from educators,
legal experts, data privacy officers and technologists to address Ethics declaration
the multifaceted nature of these issues. One potential challenge in
creating such policies is that ethical considerations are often Informed consents were obtained from all involved parties.
complex and context-dependent, making it challenging to craft
one-size-fits-all guidelines. To mitigate this, creating flexible,
Declaration of competing interest
adaptable policies with guiding principles rather than rigid rules
could be a better approach, complemented by regular revisions
The authors declare that they have no known competing financial
based on the evolving technological and societal landscape.
interests or personal relationships that could have appeared to influence
(2) Establish collaboration with stakeholders in language edu­
the work reported in this paper.
cation: Engage with governments, private organisations and
educational institutions to secure the funding, resources and
infrastructure necessary for adopting AI in English language References
teaching. This could be achieved by creating partnerships focused
Akgun, S., & Greenhow, C. (2021). Artificial intelligence in education: Addressing ethical
on AI in education and ensuring that various stakeholders are challenges in K-12 settings. AI and Ethics, 2, 431–440. https://doi.org/10.1007/
represented and involved in decision-making. However, aligning s43681-021-00096-7

6
L. Kohnke et al. Computers and Education: Artificial Intelligence 5 (2023) 100156

Ally, M. (2019). Competency profile of the digital and online teacher in future education. Lim, W. M., Gunasekara, A., Pallant, J. L., Pallant, J. I., & Pechenkina, E. (2023).
International Review of Research in Open and Distributed Learning, 20(2). https://doi. Generative ai and the future of education: Ragnarök or reformation? A paradoxical
org/10.19173/irrodl.v20i2.4206 perspective from management educators. International Journal of Management in
Anney, V. N. (2014). Ensuring the quality of the findings of qualitative research: Looking Education, 21(2), Article 100790. https://doi.org/10.1016/j.ijme.2023.100790
at trustworthiness criteria. Journal of Emerging Trends in Educational Research and Liu, M., Li, Y., Xu, W., & Liu, L. (2017). Automated essay feedback generation and its
Policy Studies, 5, 272–281. impact on revision. IEEE Transactions on Learning Technologies, 10(4), 502–513.
Bozkurt, A., Karadeniz, A., Baneres, D., Guerrero-Roldan, A. E., & Rodriguez, M. E. https://doi.org/10.1109/tlt.2016.2612659
(2021). Artificial intelligence and reflections from educational landscape: A review Luckin, R., Cukurova, M., Kent, C., & Boulay, B. (2022). Empowering educators to be AI-
of ai studies in half a century. Sustainability, 13(2), 16. https://doi.org/10.3390/ ready. Computers and Education: Artificial Intelligence, 3, Article 100076. https://doi.
su13020800 org/10.1016/j.caeai.2022.100076
Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative McCarthy, J. (2007). From here to human-level AI. Artificial Intelligence, 171(18),
Research in Psychology, 3(2), 77–101. https://doi.org/10.1191/1478088706qp063oa 1174–1182. https://doi.org/10.1016/j.artint.2007.10.009
Chaudhry, M. A., & Kazim, E. (2021). Artificial intelligence in education (AIEd): A high- McKinsey & Company. (2023). What is generative AI? Accessed Jan 20 from https
level academic and industry note 2021. AI Ethics, 2, 157–165. https://doi.org/ ://www.mckinsey.com/featured-insights/mckinsey-explainers/what-is-generative
10.1007/s43681-021-00074-z -ai.
Chen, L., Chen, P., & Lin, Z. (2020). Artificial intelligence in education: A review. IEEE Merriam, S. B., Tisdell, E. J., & E, J. (2016). Qualitative research: A guide to design and
Access, 8, 75264–75278. https://doi.org/10.1109/ACCESS.2020.2988510 implementation (4th ed.). San Francisco, CA: Jossey Bass.
Chen, X., Xie, H., Zou, D., & Hwang, G. J. (2020a). Application and theory gaps during Moorhouse, B. L. (2023). Teachers’ digital technology use after a period of online
the rise of artificial intelligence in education. Computers and Education: Artificial teaching. ELT Journal. https://doi.org/10.1093/elt/ccac050.
Intelligence, 1, 100002. https://doi.org/10.1016/j.caeai.2020.100002. Newell, A., & Simon, H. A. (1956). The logic theory machine: A complex information
Chiu, T. K. F., Xia, Q., Zhou, X., Chai, C. S., & Cheng, M. (2023). Systematic literature processing system. The Rand Corporation.
review on opportunities, challenges, and future research recommendations of Ng, D. T. K., Lee, M., Tan, R. J. Y., Hu, X., Downie, J. S., & Chu, S. K. W. (2022c).
artificial intelligence in education. Computers & Education: Artificial Intelligence, 4, A review of AI teaching and learning from 2000 to 2020. Education and Information
Article 100118. https://doi.org/10.1016/j.caeai.2022.100118 Technologies. https://doi.org/10.1007/s10639-022-11491-w
Christudas, B. C. L., Kirubakaran, E., & Thangaih, P. R. J. (2018). An evolutionary Ng, D. T. K., Leung, J. K. L., Su, J., Ng, R. C. W., & Chu, S. K. W. (2023). Teachers’ AI
approach for personalisation of content delivery in e-learning systems based on digital competencies and twenty-first century skills in the post-pandemic world.
learner behavior forcing compatibility of learning materials. Telematics and Educational Technology Research & Development, 71, 137–161. https://doi.org/
Informatics, 35(3), 520–533. https://doi.org/10.1016/j.tele.2017.02.004 10.1007/s11423-023-10203-6
Cohen, L., Manion, L., & Morrison, K. (2011). Research methods in education (7th ed.). Nigam, A., Pasricha, R., Singh, T., & Churi, P. (2021). A systematic review on AI-based
London: Routledge. proctoring systems: Past, present and future. Education and Information Technologies,
Creswell, J. W. (2008). Educational research: Planning, conducting, and evaluating 26(5), 6421–6445. https://doi.org/10.1007/s10639-021-10597-x
quantitative and qualitative research (3rd ed.). Upper Saddle River, NJ: Pearson OECD. (2022). Recommendations of the council on artificial intelligence, OECD/LEGAL/
Education, Inc. 0449. https://legalinstruments.oecd.org/en/instruments/oecd-legal-0449#:~:
Denzin, N. K., & Lincoln, Y. S. (2011). The Sage handbook of qualitative research (4th ed.). text=Download/-,Print,-Booklet. (Accessed 1 June 2023).
Thousand Oaks, CA: SAGE Publications. Papadopoulos, I., Lazzarino, R., Miah, S., Weaver, T., Thomas, B., & Koulouglioti, C.
Du, Y., & Gao, H. Determinants affecting teachers’ adoption of AI-based applications in (2020). A systematic review of the literature regarding socially assistive robots in
EFL context: An analysis of analytic hierarchy process. Education and Information pre-tertiary education. Computers & Education, 155, Article 103924. https://doi.org/
Technologies, 27, 9357-9387. https://doi.org/10.1007/s10639-022-11001-y. 10.1016/j.compedu.2020.103924
Gao, J. (2021). Exploring the feedback quality of an automated writing evaluation Popenici, S.A.D., Kerr, S. Exploring the impact of artificial intelligence on teaching and
system pigai. International Journal of Emerging Technologies in Learning, 16(11). learning in higher education. Research and Practice in Technology Enhanced
González-Calatayud, V., Prendes-Espinosa, P., & Roig-Vila, R. (2021). Artificial Learning, 12, 2. https://doi.org/10.1186/s41039-017-0062-8.
intelligence for student assessment: A systematic review. Applied Sciences, 11(12), Qian, L., Yang, Y., & Zhao, Y. (2021). Syntactic complexity revisited: Sensitivity of
5467. https://doi.org/10.3390/app11125467 China’s AES-generated scores to syntactic measures, effects of discourse-mode and
Healy, E. F., & Blade, G. (2020). Tips and tools for teaching organic synthesis online. topic. Reading and Writing, 34(3), 681–704. https://doi.org/10.1007/s11145-020-
Journal of Chemical Education, 97(9), 3163–3167. https://doi.org/10.1021/acs. 10087-5
jchemed.0c00473 Ratten, V. (2020). Coronavirus (Covid-19) and the entrepreneurship education
Hockly, N. (2023). Artificial Intelligence in English language teaching: The good, the bad community. Journal of Enterprising Communities: People and Places in the Global
and the ugly. RELC Journal. https://doi.org/10.1177/00336882231168504 Economy, 14(5), 753–764. https://doi.org/10.1108/JEC-06-2020-0121
Holmes, W., Porayska-Pomsta, K., Holstein, K., Sutherland, E., Baker, T., Shum, S. B., Sandelowsky, M. (1996). Sample size in qualitative research. Research in Nursing &
Santos, O. C., Rodrigo, M. T., Cukurova, M., Bittencourt, I. I., & Koedinger, K. R. Health, 18(2), 179–183. https://doi.org/10.1002/nur.4770180211
(2022). Ethics of AI in education: Towards a community-wide framework. Sartika, F., Ritonga, M., Lahmi, A., Rasyid, A., & Febriani, S. R. (2021). Online learning in
International Journal of Artificial Intelligence in Education, 32(3), 504–526. https://doi. the low internet area, planning, strategies and problems faced by students during the
org/10.1007/s40593-021-00239-1 Covid-19 period. Artificial intelligence for COVID-19, 413–421 (Springer).
Holmes, W., & Tuomi, I. (2022). State of the art and practice in AI in education. European Seo, K., Tang, J., Roll, I., Fels, S., & Yoon, D. (2021). The impact of artificial intelligence
Journal of Education: Research, development and policy, 57(4), 542–570. https://doi. on learner– instructor interaction in online learning. International Journal of
org/10.1111/ejed.12533 Educational Technology in Higher Education, 18(1), 1–23. https://doi.org/10.1186/
Howard, S., Tondeur, J., Hutchison, N., Scherer, R., & Siddiq, F. (2022). A t (r) opical s41239-021-00292-9
journey: Using text mining to explore teachers’ experiences in the Great Online Sharma, H., Soetan, T., Farinloye, T., Mogaji, E., & Noite, M. D. F. (2022). AI adoption in
Transition. In Society for information technology & teacher education international universities in emerging economies: Prospects, challenges and recommendations. In
conference (pp. 823–828). Association for the Advancement of Computing in E. Mogaji, V. Jain, F. Maringe, & R. E. Hinson (Eds.), Re-Imagining educational futures
Education (AACE). in developing countries. Cham: Palgrave Macmillan. https://doi.org/10.1007/978-3-
Hrastinski, S., Olofsson, A. D., Arkenback, C., et al. (2019). Critical imaginaries and 030-88234-1_9.
reflections on artificial intelligence and robots in postdigital K-12 education. Sijing, L., & Lan, W. (2018). Artificial intelligence education ethical problems and
Postdigital Science Education, 1, 427–445. https://doi.org/10.1007/s42438-019- solutions. In 2018 13th international conference on computer science & education
00046-x (ICCSE) (pp. 1–5). IEEE.
Hwang, G. J., Xie, H., Wah, B. W., & Gašević, D. (2020). Vision, challenges, roles and Torda, A. (2020). How COVID-19 has pushed us into a medical education revolution.
research issues of artificial intelligence in education. Computers and Education: Internal Medicine Journal, 50(9), 1150–1153, 10.1111%2Fimj.14882.
Artificial Intelligence, 1, Article 100001. https://doi.org/10.1016/j. Weng, X., & Chiu, T. K. F. (2023). Instructional design and learning outcomes of
caeai.2020.100001 intelligent computer assisted language learning: Systematic review in the field.
Kabudi, T., Pappas, I., & Olsen, D. H. (2021). AI-enabled adaptive learning systems: A Computers and Education: Artificial Intelligence, 4. https://doi.org/10.1016/j.
systematic mapping of the literature. Computers and Education: Artificial Intelligence, caeai.2022.100117
2, Article 100017. https://doi.org/10.1016/j.caeai.2021.100017 Whalley, B., France, D., Park, J., Mauchline, A., & Welsh, K. (2021). Towards flexible
Karaca, O., Çalıs ķan, S. A., & Demir, K. (2021). Medical artificial intelligence readiness personalized learning and the future educational system in the fourth industrial
scale for medical students (MAIRS-MS)–development, validity and reliability study. revolution in the wake of Covid- 19. Higher Education Pedagogies, 6(1), 79–99.
BMC Medical Education, 21(1), 1–9. https://doi.org/10.1186/s12909-021-02546-6 https://doi.org/10.1080/23752696.2021.1883458
Kexin, L., Yi, Q., Xiaoou, S., & Yan, L. (2020). Future education trend learned from the Williamson, B., & Eynon, R. (2020). Historical threads, missing links, and future
Covid-19 pandemic: Take artificial intelligence online course as an example. In 2020 directions in AI in education. Learning, Media and Technology, 45(3), 223–235.
international conference on artificial intelligence and education (ICAIE) (pp. 108–111). https://doi.org/10.1080/17439884.2020.1798995
IEEE. Xia, Q., Chiu, T. K., Lee, M., Sanusi, I. T., Dai, Y., & Chai, C. S. (2022). A self-
Kohnke, L., Moorhouse, B. L., & Zou, D. (2023). ChatGPT for Language Teaching and determination theory (SDT) design approach for inclusive and diverse artificial
Learning. RELC Journal. https://doi.org/10.1177/00336882231162868.
Kong, S. C., Cheung, W. M. Y., & Zhang, G. (2021). Evaluation of an artificial intelligence
literacy course for university students with diverse study backgrounds. Computers
and Education: Artificial Intelligence, 2, Article 100026. https://doi.org/10.1016/j.
caeai.2021.100026

7
L. Kohnke et al. Computers and Education: Artificial Intelligence 5 (2023) 100156

intelligence (AI) education. Computers & Education, 189, Article 104582. https://doi. ://www.scmp.com/news/hong-kong/education/article/3210650/university-hong-
org/10.1016/j.compedu.2022.104582 kong-temporarily-bans-students-using-chatgpt-other-ai-based-tools-coursework.
Yau, C., & Chan, K. (2023). University of Hong Kong temporarily bans students from Zhang, K., & Aslan, A. B. (2021). AI technologies for education: Recent research & future
using ChatGPT, other AI-based tools for courseworks. South China Morning Post. https directions. Computers and Education: Artificial Intelligence, 2, Article 100025. https://
doi.org/10.1016/j.caeai.2021.100025