EU-backed tool to combat cheating in online exams finally ready

‘Tesla’ creators say accuracy and security in design were more important than rushed roll-out

四月 27, 2021
Group of people in same costume wearing same smiling mask as a metaphor for a tool using facial and keystroke recognition has been launched.
Source: Getty

A European Union-backed tool designed to stamp out cheating in online assessment with facial and keystroke recognition is finally ready for launch and will be available for free.

Despite having reported successful pilots in 2018, the €7 million (£6 million) Adaptive Trust-based E-assessment System for Learning (Tesla) tool had not made it to market by the time the coronavirus pandemic forced universities to switch to remote assessment last year.

This was because the researchers wanted to make sure that the technology was effective and worked across a range of disciplines, and that the handling of personal information would comply with data protection laws, explained Ana Elena Guerrero Roldán, who coordinates the Tesla project at the Open University of Catalonia.


THE Campus resource: Fair assessment and tackling the rise in online cheating


The team has created a new tool, Tesla Community Edition, that uses facial and keystroke recognition to check that online exam candidates are who they say they are and are not cheating. It will also compare learners’ work to check for potential plagiarism.

The community edition will be free and should be available for use by this summer.

Amid reports of a spike in cheating, technology companies rushed out online proctoring products during 2020. But many faced a backlash from students who found tools such as video monitoring intrusive, while academics questioned the efficacy of such products.

Denise Whitelock, professor of technology-enhanced assessment and learning at the UK’s Open University, which was part of the EU project, said universities should be reassured that the EU had helped to develop Tesla. All Tesla data and communications are encrypted, and the team behind the tool has produced a series of peer-reviewed journal papers.

“We have had lawyers working with us from the beginning to ensure that the way the data is handled within the system complies with the law, and this was built into its architecture,” Professor Whitelock said.

More than 27,000 students at seven European universities took part in trials of Tesla. This led to the removal of a “forensic analysis” tool, which compared writing styles to check authorship, because the results were not clear enough.

For Dr Guerrero Roldán, the key advantage of Tesla is that it was designed to support continuous assessment rather than just catch cheaters.

“If you, as an institution, think only about cheating, then you are not devoting efforts on rethinking your educational model according to modern times,” she said. “If institutions are able to track learners across the whole learning path in a course, then final examinations can be reduced because authentication and authorship can be assured across the whole learning process.”

Professor Whitelock rejected any notion that Tesla had missed its big opportunity when Covid hit, emphasising that the team had been “testing and testing this product for a long time”.

“We’re now ready for the next phase [of online assessment], and with everyone concerned about the need for quality assurance, it is a great time for Tesla,” she said.

anna.mckie@timeshighereducation.com

请先注册再继续

为何要注册?

  • 注册是免费的,而且十分便捷
  • 注册成功后,您每月可免费阅读3篇文章
  • 订阅我们的邮件
注册
Please 登录 or 注册 to read this article.

相关文章

Reader's comments (4)

A good Lecturer should be able to ascertain whether a student is being fraudulent in his/her work from their knowledge of the student as well as a comparison of their previous work. I found that 2 students submitted the same essay and I failed both of them - however my Head of School told me that these students are paying good money to stay at University, and they should be passed - I refused and was abused, harassed, and discriminated against and was SACKED from my Tenured Lectureship because I refused to pass students who cheated. Australian Universities are not as honest as they portray.
A good lecture usually does in theory and I am rather critical of this development akin to surveillance too. However, try this with a cohort of 200 plus students on a course/module and anonymous marking. Good luck with knowing every single student well enough to make that call or having the time to probe plagiarism without technical tools. Not to mention marking a student's work in 10-15 minutes per piece max (workload allowance for marking, which usually includes feedback and admin too). Welcome to the realities of the modern sausage factory, sorry university.
It is a matter for wry amusement that a snooping tool has the word 'trust' as part of its name! This is all about not trusting students but surveilling them. Instead, write better assessments that require them to use their knowledge, to apply it. They won't find the answers ready-made... and even if they do, think about how we work in the real world. We talk to colleagues, we google (or duckduckgo), we read books, we find YouTube tutorials... the one thing we rarely do is rely on what we can remember, even though the basic knowledge we possess is being used to assess and evaluate what we have found elsewhere, deciding how and if it can help us deal with whatever issue is to hand.
Agree with "An Academic Somewhere" - I too run a course with 200 students - but it is pretty easy to spot plagiarism that the tech does not because poorly engaged students copy and paste from scientific articles then use a synonym finder to change just enough words to "beat" the plagiarism detection tech, often rendering their text meaningless, because there are very few true synonyms in English. For example "right hemisphere of the brain" becomes "correct meridian of the mind". Poetic, but nonsensical. I routinely report this behaviour but instead of dealing with the issue of student plagiarism, my University has told me to redesign my assessments - not straightforward in science where students are required to critically evaluate previously published work. Hard to do this without students reading (and in some cases copying from) the literature.