Can computers really mark exams? Benefits of ELT automated assessments

ɫèAV Languages
Hands typing at a laptop with symbols

Automated assessment, including the use of Artificial Intelligence (AI), is one of the latest education tech solutions. It speeds up exam marking times, removes human biases, and is as accurate and at least as reliable as human examiners. As innovations go, this one is a real game-changer for teachers and students. 

However, it has understandably been met with many questions and sometimes skepticism in the ELT community – can computers really mark speaking and writing exams accurately? 

The answer is a resounding yes. Students from all parts of the world already take AI-graded tests.  aԻ Versanttests – for example – provide unbiased, fair and fast automated scoring for speaking and writing exams – irrespective of where the test takers live, or what their accent or gender is. 

This article will explain the main processes involved in AI automated scoring and make the point that AI technologies are built on the foundations of consistent expert human judgments. So, let’s clear up the confusion around automated scoring and AI and look into how it can help teachers and students alike. 

AI versus traditional automated scoring

First of all, let’s distinguish between traditional automated scoring and AI. When we talk about automated scoring, generally, we mean scoring items that are either multiple-choice or cloze items. You may have to reorder sentences, choose from a drop-down list, insert a missing word- that sort of thing. These question types are designed to test particular skills and automated scoring ensures that they can be marked quickly and accurately every time.

While automatically scored items like these can be used to assess receptive skills such as listening and reading comprehension, they cannot mark the productive skills of writing and speaking. Every student's response in writing and speaking items will be different, so how can computers mark them?

This is where AI comes in. 

We hear a lot about how AI is increasingly being used in areas where there is a need to deal with large amounts of unstructured data, effectively and 100% accurately – like in medical diagnostics, for example. In language testing, AI uses specialized computer software to grade written and oral tests. 

How AI is used to score speaking exams

The first step is to build an acoustic model for each language that can recognize speech and convert it into waveforms and text. While this technology used to be very unusual, most of our smartphones can do this now. 

These acoustic models are then trained to score every single prompt or item on a test. We do this by using human expert raters to score the items first, using double marking. They score hundreds of oral responses for each item, and these ‘Standards’ are then used to train the engine. 

Next, we validate the trained engine by feeding in many more human-marked items, and check that the machine scores are very highly correlated to the human scores. If this doesn’t happen for any item, we remove it, as it must match the standard set by human markers. We expect a correlation of between .95-.99. That means that tests will be marked between 95-99% exactly the same as human-marked samples. 

This is incredibly high compared to the reliability of human-marked speaking tests. In essence, we use a group of highly expert human raters to train the AI engine, and then their standard is replicated time after time.  

How AI is used to score writing exams

Our AI writing scoring uses a technology called . LSA is a natural language processing technique that can analyze and score writing, based on the meaning behind words – and not just their superficial characteristics. 

Similarly to our speech recognition acoustic models, we first establish a language-specific text recognition model. We feed a large amount of text into the system, and LSA uses artificial intelligence to learn the patterns of how words relate to each other and are used in, for example, the English language. 

Once the language model has been established, we train the engine to score every written item on a test. As in speaking items, we do this by using human expert raters to score the items first, using double marking. They score many hundreds of written responses for each item, and these ‘Standards’ are then used to train the engine. We then validate the trained engine by feeding in many more human-marked items, and check that the machine scores are very highly correlated to the human scores. 

The benchmark is always the expert human scores. If our AI system doesn’t closely match the scores given by human markers, we remove the item, as it is essential to match the standard set by human markers.

AI’s ability to mark multiple traits 

One of the challenges human markers face in scoring speaking and written items is assessing many traits on a single item. For example, when assessing and scoring speaking, they may need to give separate scores for content, fluency and pronunciation. 

In written responses, markers may need to score a piece of writing for vocabulary, style and grammar. Effectively, they may need to mark every single item at least three times, maybe more. However, once we have trained the AI systems on every trait score in speaking and writing, they can then mark items on any number of traits instantaneously – and without error. 

AI’s lack of bias

A fundamental premise for any test is that no advantage or disadvantage should be given to any candidate. In other words, there should be no positive or negative bias. This can be very difficult to achieve in human-marked speaking and written assessments. In fact, candidates often feel they may have received a different score if someone else had heard them or read their work.

Our AI systems eradicate the issue of bias. This is done by ensuring our speaking and writing AI systems are trained on an extensive range of human accents and writing types. 

We don’t want perfect native-speaking accents or writing styles to train our engines. We use representative non-native samples from across the world. When we initially set up our AI systems for speaking and writing scoring, we trialed our items and trained our engines using millions of student responses. We continue to do this now as new items are developed.

The benefits of AI automated assessment

There is nothing wrong with hand-marking homework tests and exams. In fact, it is essential for teachers to get to know their students and provide personal feedback and advice. However, manually correcting hundreds of tests, daily or weekly, can be repetitive, time-consuming, not always reliable and takes time away from working alongside students in the classroom. The use of AI in formative and summative assessments can increase assessed practice time for students and reduce the marking load for teachers.

Language learning takes time, lots of time to progress to high levels of proficiency. The blended use of AI can:

  • address the increasing importance of formative assessmentto drive personalized learning and diagnostic assessment feedback 

  • allow students to practice and get instant feedback inside and outside of allocated teaching time

  • address the issue of teacher workload

  • create a virtuous combination between humans and machines, taking advantage of what humans do best and what machines do best. 

  • provide fair, fast and unbiased summative assessment scores in high-stakes testing.

We hope this article has answered a few burning questions about how AI is used to assess speaking and writing in our language tests. An interesting quote from Fei-Fei Li, Chief scientist at Google and Stanford Professor describes AI like this:

“I often tell my students not to be misled by the name ‘artificial intelligence’ — there is nothing artificial about it; A.I. is made by humans, intended to behave [like] humans and, ultimately, to impact human lives and human society.”

AI in formative and summative assessments will never replace the role of teachers. AI will support teachers, provide endless opportunities for students to improve, and provide a solution to slow, unreliable and often unfair high-stakes assessments.

Examples of AI assessments in ELT

At ɫèAV, we have developed a range of assessments using AI technology.

Versant

The Versant tests are a great tool to help establish language proficiency benchmarks in any school, organization or business. They are specifically designed for placement tests to determine the appropriate level for the learner.

PTE Academic

The  is aimed at those who need to prove their level of English for a university place, a job or a visa. It uses AI to score tests and results are available within five days. 

ɫèAV English International Certificate (PEIC)

ɫèAV English International Certificate (PEIC) also uses automated assessment technology. With a two-hour test available on-demand to take at home or at school (or at a secure test center). Using a combination of advanced speech recognition and exam grading technology and the expertise of professional ELT exam markers worldwide, our patented software can measure English language ability.

Read more about the use of AI in our learning and testing here, or if you're wondering which English test is right for your students make sure to check out our post 'Which exam is right for my students?'.

Explore our courseware and exams

More blogs from ɫèAV

  • A girl sat at a laptop with headphones on in a library

    5 myths about online language learning

    By Steffanie Zazulak
    Reading time: 3 minutes

    Technology has radically changed the way people are able to access information and learn. As a result, there are a great number of tools to facilitate online language learning – an area that’s been the subject of many myths. Here we highlight (and debunk) some of the bigger ones…

    Myth #1: You will learn more quickly

    Although online learning tools are designed to provide ways to teach and support the learner, they won’t provide you with a shortcut to proficiency or bypass any of the key stages of learning.Although you may well be absorbing lots of vocabulary and grammar rules while studying in isolation, this isn’t a replacement for an environment in which you can immerse yourself in the language with English speakers. Such settings help you improve your speaking and listening skills and increase precision, because the key is to find opportunities to practise both – widening your use of the language rather than simply building up your knowledge of it.

    Myth #2: It replaces learning in the classroom

    With big data and AI increasingly providing a more accurate idea of their level, as well as a quantifiable idea of how much they need to learn to advance to the next level of proficiency, classroom learning is vital for supplementing classroom learning. And with the Global Scale of English providing an accurate measurement of progress, students can personalise their learning and decide how they’re going to divide their time between classroom learning and private study.

    Myth #3: It can’t be incorporated into classroom learning

    There are a huge number of ways that students and teachers can use the Internet in the classroom. Meanwhile, ɫèAV’s online courses and apps have a positive, measurable impact on your learning outcomes.

    Myth #4:You can't learn in the workplace

    Online language learning is ideally suited to the workplace and we must create the need to use the language and opportunities to practise it. A job offers one of the most effective learning environments: where communication is key and you’re frequently exposed to specialized vocabulary. Online language learning tools can flexibly support your busy schedule.

    Myth #5: Online language learning is impersonal and isolating

    A common misconception is that online language learning is a solitary journey, lacking the personal connection and support found in traditional classrooms. In reality, today’s digital platforms are designed to foster community and real interaction. With features like live virtual classrooms, discussion forums and instant feedback, learners can connect with peers and educators around the world, building skills together.

  • Two teenagers sat at a desk in a classroom working together in front of a laptop

    My lifelong learning journey: Why learning English never stops

    By Zarela Cruz
    Reading time: 4 minutes

    My journey with English began in the unlikeliest of places: a mining camp in southern Peru. As a child, I was fascinated by American culture – the movies, the music, the seemingly limitless world that English opened up. For me, the language was a gateway leading to a deeper understanding and feeling of belonging, making me part of their culture.

  • A classroom scene with a teacher and diverse students engaged in learning, using laptops on desks, in a brightly lit room.

    Is game-based learning technology a waste of time?

    By
    Reading time: 4 minutes

    We feel that game-based learning (GBL) is a waste of time… if not properly understood. Even then, one could argue that "wasting time" is a vital part of learning and perhaps we need to stop insisting that every second counts.

    Game-based learning vs. Gamification: Understanding the difference

    Let’s begin by first addressing the term "game-based learning" and how it compares to its doppelgänger, gamification. Gamification is the application of game mechanics and dynamics to non-game contexts to solve problems, engage users and promote desired behaviours. For example, rewarding acheivements with points, awards or badges for achievement, levelling up, using avatars, quests and collaboration are all gamification features, which elevate games above the mundane activities of normal life.

    The principles of gamification have been applied to the retail and services sectors for years: think of airlines' frequent flyers programs, pubs and bars running "happy hours" promotions, WeightWatchers' points-counting, Foursquare's badges for visiting new places. There are hundreds of examples of where game dynamics have been introduced into non-game contexts to influence behaviour and bring about a desired result.

    Game-based learning in action: More than just play

    Gamification does not refer to the straightforward use of games (whether digital or otherwise) as part of a teaching or learning interaction. A teacher using the board game Monopoly in the classroom to demonstrate the idea of rent is not gamifying the learning environment; they are involving learners in game-based learning. GBL refers to the use of games as tools: as devices for opening discussion, presenting concepts or promoting learner engagement within clearly defined learning objectives, in other words, learning through playing games. Gamification is the appropriation of those principles, mechanics and dynamics that make games work in order to promote engagement or engender a desired outcome.

    Teachers have long introduced games into the learning environment – such as Kim’s Game, Pelmanism and Guess Who? – and the uptake of GBL with digital games is particularly well demonstrated by the work being done by the Institute of Play, and the growing popularity of the likes of Minecraft and SimCity in schools. In terms of our immediate ELT context, pretty much any video game can be repurposed for language learning, in the same way a text, song or website can be. Take, for example, the indie game . Although there is no actual spoken language in the game, the platform puzzle format lends itself perfectly to practising language around predictions or conditionals ("If I pull that lever, the door will open"), recounting events ("I was chased by a giant spider!"), strategising and so on. In this post, we will be referring to existing digital games that have been appropriated into a learning context, as opposed to games that have been designed with a specific educational use in mind.

    Sharma and Barrett’s definition of blended learning provides a useful context for approaching the use of digital games in such modalities, notably the combination of “a face-to-face classroom component with an appropriate use of technology”. We’re making no assumptions about whether the games are being accessed in the classroom, on mobile or online at home. The blend isn’t defined by where a learner is, but by how their use of technology supports and enhances their contact with the teacher.

    The notion of appropriateness in Sharma and Barrett’s definition is critical as, in the case of a GBL project, it assumes a teacher has a familiarity with both the tech and content accessibility of a selected game. Acquiring that level of familiarity with a game requires a certain amount of time engaging with it to determine its fit for the needs of the learners, an activity which might easily be considered off task when compared to the other demands being made on an educator’s schedule.

    How GBL fuels engagement and deeper learning

    The benefits of GBL with digital games are potentially quite profound, however. First, studies indicate that playing video games in general can stimulate the generation of neurons and enhance connectivity between the regions of the brain responsible for memory formation, spatial orientation and strategic thinking. The right pairing of game and learning objectives could be argued to promote situated cognition, a theory that knowledge is constructed through – and inseparable from – social interactions and the context in which they take place. A learner immersed in SimCity stands a much greater chance of understanding the principles of taxation and the provision of public services through playing the role of a mayor, for example, than a learner being walked through the annual budget. As Lim et al. state: “games are effective because learning takes place within a meaningful context where what must be learned is directly related to the environment in which learning and demonstration take place”.

    Early-stage research on mirror neurons is adding a new aspect to the discussion around the immediacy of playing games. In short, mirror neurons suggest that when we observe someone performing an action, there is a brief moment in which our brain cells fire as if we are carrying out that action ourselves. The boundary between observer (player) and observed (in-game character) becomes blurred for a split second. A lot can happen in that split second.

    In addition to the benefits of an immersive, neuron-stroking experience, games demonstrably promote learner engagement by introducing the F-bomb into the mix (fun). They also have the capacity to provide an unrivalled social experience, as in the case of MMORPGs (massively multiplayer online role-playing games). As an example, consider how a leading MMORPG game, World of Warcraft, is applied in learning environments.

    Addressing concerns and embracing change

    However, there are often deep-seated cultural aversions to the use of games in an educational context that GBL initiatives are required to overcome. Critics have said that digital games are anti-social, that they rot your attention span, that they are not legitimate, validated learning resources. Although there is not yet a body of research that can empirically confirm or debunk the effectiveness of games used for learning, surely watching a learner plan, execute and evaluate a project in Minecraft with classmates suggests that those objections are based on dated assumptions. The language learning space in particular is still very much attached to a coursebook paradigm that is predicated on levels and a clearly defined syllabus. Perhaps GBL is too much at odds with an established business model that is the bedrock of too many large education organisations.

    So is GBL a waste of time? We’d argue that it is when its potential is not properly recognised and it is treated as light relief. Games are dynamic, engaging resources capable of delivering experiences and drawing connections that can really ignite a student’s learning experience. Furthermore, they bring a playful and unpredictable aspect to the learning process.