Introducing AI and what students think about it

As artificial intelligence (AI) becomes increasingly available, it's essential to understand how students are engaging with these tools. At UQ, over 80% of students report using AI in their academic studies. By proactively addressing AI's role in learning, we can help students navigate its opportunities and challenges responsibly.

What is artificial intelligence (AI), generative AI (GenAI), and machine translation (MT)?

Artificial intelligence (AI)

Artificial intelligence (AI) is embedded into our digital world. AI systems simulate intelligence to perform specific tasks, like recommending products, transcription, or recognising faces.

You likely use AI every day. Examples include:

grammar checkers: Grammarly, Microsoft Word’s editor tools
spelling and autocorrect tools: Built into word processors and browsers
search engines: Google, Bing
voice-to-text transcriptions: Apple Dictation, Zoom audio transcription, Word transcription
data analysis tools: Excel functions, SPSS, MATLAB.

Machine Translation (MT)

Machine Translation (MT) is a type of artificial intelligence that serves to translate text and/or speech from one language into another.

Examples of tasks that use MT include:

converting an assessment that you wrote from one language to another
translating a website into another language to help you to interpret it
interpreting live speech broadcasts.

Generative AI (GenAI)

Generative AI (GenAI) is a newer type of AI that creates new things (based on patterns in existing data) in response to user prompts. It can be used to make images, text, music or videos.

Tools like ChatGPT and Microsoft Copilot are examples of GenAI. Existing tools like Microsoft PowerPoint are increasingly including GenAI to generate slide decks or images from your text or voice prompts, for example.

Use of the term 'AI'

‘AI’ is typically used to refer to all AI, including GenAI and MT.

AI is evolving rapidly, and it's not always clear when a tool is using AI or GenAI. For example, Grammarly is well known for improving grammar in writing, however, it also translates written text in over 15 languages, and some versions generate text. ChatGPT can also be used to refine grammar and translate text from one language to another.

Further resources

For more in-depth definitions of AI, read:

How students use and feel about AI

UQ students shared their perspectives in 2024 and 206 through the AI in Higher Education survey.

What has changed since 2024?

AI use is becoming more regular across UQ students, with weekly or daily use rising from 41% to 57%. Students’ motivations are becoming more pragmatic: ease (+8%) and speed (+6%) gained ground, while enjoyment declined (−6%). Fewer practical barriers suggest students are becoming more familiar with AI and better able to access tools, alongside stronger principled avoidance: more students cite ethical concerns (+9%) and a preference for doing the work themselves (+8%).

What stands out in 2026?

Confidence and Perceptions of AI

Overall, student engagement with AI and confidence in using AI for study were strongest in BEL and EAIT and lowest in HASS (confidence: 62%, 55% and 27%, respectively), while increased use since 2024 was accompanied by declining trust in AI outputs. Most students agreed that assessment should avoid AI dependence (74%), but views on whether allowing AI devalues a degree differed considerably across faculties (32%–64%).

AI in Assessment

Notably, students perceived AI-restricted assessments as better supporting deep learning (66% vs. 48%) and demonstrating authentic capability (69% vs. 38%) than assessments that allowed AI. Moreover, faculty differences were more pronounced when AI was allowed, particularly in perceptions of producing high-quality work, learning deeply and demonstrating capability.

These disciplinary contrasts also extend to students’ reasons for using and avoiding AI. Students across disciplines share similar practical barriers, but differ more on principle. HASS students more often report a preference for doing the work themselves, with a 33% gap between disciplines. Ethical concerns show an even larger gap (36%), again highest in HASS. Meanwhile, BEL students more often identify faster study as a motivation, with a 36% between disciplines.

Guidance and support

Guidance and preparation remain significant gaps. Although strong concern about breaking UQ’s AI rules fell from 61% in 2024 to 46% in 2026, 84% of students still expressed some concern. Only 35% felt supported to use AI effectively in their studies, and 26% felt prepared to use it in their future professions, up from 24% and 17%, respectively.

Prioritising Critical AI Use

These findings suggest supporting students to use AI critically, not simply more often. Priorities include discipline-specific guidance, assessment that makes independent capability visible, and respect for students who choose to limit or avoid AI.

Read the full report (PDF, 1 MB) on students' responses to the AIinHE survey.

AI weaknesses and risks for learning

Students are aware of the risks to their skill development if reliant on AI for tasks essential to learning (critical thinking, problem solving, research analysis), and want to retain the integrity of their learning experience. - Student Voice Forum, July 2024

As a university, we need to guide students in using AI responsibly and effectively for their study (Lodge et. al., 2025). While AI tools can enhance learning, they also have significant limitations:

Inaccuracies: AI models may have a tendency to blend or fabricate information.
Lack of source attribution: Content from AI may not credit its original sources.
Biases: Any biases present in the data used to train AI can become amplified in the model’s outputs.
Limitations in human-like thinking: AI tools simulate intelligence but do not replicate the depth and nuance of human cognition.
Data privacy concerns: Most AI tools have license agreements granting them access to any data you input. This includes student information and assessments, which must be protected.
Overreliance: Students may become overly reliant on AI, which could potentially hinder their learning by offloading cognitive activities important for learning.

While many concerns about AI, student usage, and its overall impact are unverified, leading to further confusion in this space, a rigorous, collaborative and investigative approach is underway to help understanding AI’s capabilities and limitations for learning (e.g. Lodge & Loble's (2026) exploration of cognitive offloading, Fitgerald's (2025) collection of case studies and broader investigations supported by UQ's Teaching Innovation Grants).

By being well informed about the use of AI, we can help students use these tools effectively while maintaining academic integrity and protecting learner autonomy.