International student admissions in the age of AI: why test integrity matters

Selecting the right English proficiency and admissions test for international student applicants can influence whether a school offers a place, the student’s year level and how much English preparation they complete before entering mainstream classes.

If the result overstates their abilities, the student may arrive with learning needs that neither the family nor the school has anticipated. As AI and other forms of assistance become increasingly accessible, admissions teams need confidence in both the assessment and the conditions under which it was completed.

Convenient access and fast results matter. So does knowing that the result reflects what the student can do independently.

Can the test provider properly control the testing environment?

Remote assessments may use identity checks, multiple cameras and live supervision. These measures can improve visibility and deter misconduct, but their effectiveness depends on what supervisors can see, recognise and control.

A wider camera view does not establish that every source of assistance has been excluded. How does the provider prevent access to translation tools, AI, concealed devices or another person’s help? Does supervision occur live, allowing intervention, or through recordings reviewed afterwards?

Research gives schools reason to scrutinise security claims. A 2022 USENIX study found that anti-cheating controls in four remote proctoring systems could be bypassed. Much of the available body of current research pertains to higher education and particular proctoring systems, rather than school admissions tests. It supports asking for evidence about the specific assessment a school accepts.

School examinations offer a relevant example too. The UK’s Joint Council for Qualifications documented a GCSE case in which a student accessed AI because internet access had not been disabled on their word processor. Supervision did not compensate for a failure to control the available technology.

AEAS combines direct supervision with practical control of the room. Trained Test Administrators check identity against original photographic identification, manage entry and enforce restrictions on belongings and devices.

AEAS’s smart device policy covers devices capable of communicating, recording, storing information or providing translation or AI assistance. This includes smart watches, glasses, rings and earbuds. The policy addresses capability, rather than relying on a list of particular brands or models.

Prohibited devices must be switched off and kept outside the test room. Possession inside the room breaches the rules, even if the device is not used. Checks include respectful visual inspection of glasses and, where used, non-contact wand screening. Ordinary prescription glasses remain permitted.

If a prohibited device is found, the student is removed from the room and the incident documented for review. Consequences can include withholding or cancelling results and restrictions on future testing.

These measures reduce opportunities for assistance before it occurs, while enabling direct intervention during the assessment.

What safeguards protect the child in an online environment?

A testing provider is responsible for its own conduct, but that does not remove a school’s separate child safety responsibilities. Schools must consider the child protection laws and standards applicable in their state or territory when engaging external services. Where a school requires or refers an applicant for testing, it should establish how that arrangement fits within its safeguarding obligations.

This makes provider selection more than a question of test security. What evidence demonstrates compliance with applicable child safety requirements? What written warranties does the provider give about staff screening, supervision, safeguarding and the handling of children’s information? Do those assurances also cover subcontractors and invigilators based overseas?

When a child takes a remotely supervised test at home, an invigilator unknown to the family may observe and record their bedroom or another private space. Schools should know how that person is checked, trained and supervised, how they communicate with children and what happens if a child becomes distressed or needs help.

Recording arrangements also need clear answers: what is captured, where it is stored, who can access it, with whom it is shared and when it is deleted. Providers should explain whether recordings can be used for other purposes, including developing AI systems. Written assurances should be supported by evidence of the procedures in place.

All AEAS Test Administrators undergo stringent screening, including police checks or Working with Children Checks, and are subject to ongoing training, quality assurance and auditing. Testing takes place in person, with an administrator available to respond directly, and does not require remote video observation of the child’s home.

Could technology affect the student’s results?

Using an iPad or smartphone does not establish that a child can independently manage a formal assessment using a keyboard, mouse and unfamiliar interface.

Where testing is offered to children as young as four or five, schools should ask how its format, duration and attention demands have been evaluated. Providers should also explain how they distinguish limited digital skills from limited English proficiency, and what help adults may provide without influencing answers.

For older students, digital confidence raises a different question: are the provider’s safeguards keeping pace with the workarounds a determined candidate could find and use? A student need not develop a sophisticated bypass themselves if instructions or tools are available to them. This does not imply that teenagers generally cheat. It is, however, worth considering the high-pressure environment in which many applicants take these tests, with expectations from parents or education agents and access to a preferred school dependent on reaching an entry benchmark. How regularly are the provider’s security controls tested and updated against emerging methods of unauthorised assistance?

Connectivity creates another potential obstacle. Applicants’ access to reliable internet and overseas services varies, including in markets such as China and Iran. An online test may require the question platform, audio, multiple video feeds and uploads to function simultaneously.

What happens if listening audio cuts out, an answer fails to upload or the invigilator’s camera feed disconnects? Schools need clear procedures for restoring time, identifying affected results and arranging fair retesting.

AEAS uses age-appropriate content, written tasks and direct interaction with an examiner. Students do not need to navigate an online testing platform or maintain a home internet connection throughout the assessment. Its in-person supervision and device restrictions also reduce reliance on software identifying every possible digital workaround.

Is supervision truly independent, and are concerns investigated?

The person supervising a test should be independent of the student’s placement outcome. Where an education agent receives commission for placing a student, asking that agent to supervise their assessment creates a conflict of interest. Moving the test to an office does not resolve that question.

AEAS appoints, trains and monitors its Test Administrators independently of education agents. Agents may assist with registration, while responsibility for administering the assessment rests with AEAS.

Accountability also matters when answers are marked and reports prepared. Algorithms and AI systems can make errors, and generative AI can produce confident explanations containing unsupported assumptions or factually incorrect statements. Automated scoring is not inherently unreliable, but schools should ask how it has been validated for the students being assessed, who reviews questionable results and how mistakes are corrected.

The quality of feedback deserves similar attention. Where automated marking selects boilerplate comments from a rubric or score band, a report may describe a performance category without explaining the individual student’s difficulties. “Needs to improve vocabulary”, for example, gives a family little direction about which aspects require attention or how those limitations affect communication. An apparently personalised report is only useful if its comments are grounded in that student’s actual performance.

AEAS’s in-person assessment and live interview provide a basis for individual, insightful reporting. Direct observation and follow-up questions allow examiners to explore how a student understands instructions, develops answers and communicates spontaneously. Those observations support personalised comments about behaviour, strengths and areas for improvement, giving schools and families more useful direction than a score accompanied by generic rubric statements.

Human judgement also requires safeguards. AEAS Reports undergo internal quality assurance, and student results are monitored closely for inconsistencies across skills, previous results, learning history and interview information. An unusual result is a reason for investigation, rather than proof of misconduct.

Where concerns arise, AEAS may require a further interview or retesting and withhold a report pending review. Results may be cancelled where sufficient evidence establishes an issue affecting their validity.

Reports are sent directly to nominated schools when released to families and nominated agents, reducing reliance on documents forwarded by applicants.

Why these questions matter now

Education bodies in Australia and around the world are rapidly being forced to reconsider how technology affects student learning and assessment. NSW’s restrictions on take-home HSC assessments seek to protect evidence of students’ own abilities. Its school screen-use limits and South Australia’s broader Royal Commission into AI reflect increasing attention at home as to where technology adds value and where better safeguards are needed.

For admissions teams, the practical implication is to apply the same scrutiny to externally delivered tests. Online access, automated marking and rapid reporting are enticing delivery features, but schools still need evidence that whichever assessment they choose accurately measures independent ability, suits the student’s age and produces reliable information for entry decisions.

Before selecting a test, ensure you undertake due diligence. Interrogate whether the testing provider can demonstrate how unauthorised assistance is prevented, how questionable or unusual results are reviewed, and how its recommendations help determine English preparation and placement? If those answers are unclear, convenience alone is an insufficient basis for trusting the result.

AEAS’s tried and trusted, in-person approach gives schools concrete safeguards to assess: direct control of the testing environment, independently appointed administrators, live interaction with the student and individual reporting. These are practical reasons to choose AEAS when the result will help determine a young person’s school pathway.

Does the report help your school plan entry?

A secure result must also answer the questions that prompted your school to request an assessment in the first place: which year level is appropriate, what English preparation does the student need and when can they realistically enter mainstream classes?

A percentage or standalone score is only useful if your admissions team understands what it measures. How is it calculated for each skill? Does it represent correct answers, language proficiency or a prediction of future performance? What evidence connects it to the demands of your school’s classrooms?

Schools should expect clear explanations of these measures and their limitations. If a report cannot translate its scores into meaningful guidance about placement and preparation, it may result in assumptions being made at precisely the point where accurate and informed advice is crucial.

The AEAS Report makes data-driven recommendations about the appropriate year level and timing of entry, together with a minimum period of intensive English language preparation (ELICOS). That preparation recommendation, expressed in weeks, gives schools and families a practical basis for establishing the student’s pathway and likely entry point.

These recommendations draw on assessment of reading, vocabulary, writing, listening and speaking, alongside mathematical reasoning and non-verbal general ability. Results are considered together with educational background, school reports, interview information and observations during testing.

For admissions staff, this is information and guidance they know they can trust and rely upon when making an offer and planning commencement. For families, it makes the preparation pathway clearer. For teaching staff, the individual skill results and comments help identify the personalised support the student may need.

For Australian schools, using AEAS Testing absolutely fulfils the regulatory requirements to have a documented assessment process as required under Standard 2 of the National Code.

AEAS believes best practice requires safeguards before, during and after testing: verified identity, a controlled environment, independent supervision, restrictions on unauthorised technology, investigation of concerns and dependable reporting.

The value is a result your school can trust and recommendations it can act on, helping the student enter at an appropriate level, at an appropriate time and with the English preparation they need.

To discuss AEAS Testing, school entry benchmarks or assessment arrangements, contact [email protected] or +61 3 9645 0077.