Rethinking assessments with tech: What needs to change in the AI era?

Artificial intelligence is already impacting the way students complete assessments. 

According to a 2023 study by YouthInsight, nearly one in two (42%) of Australian students aged 14 to 17 have used ChatGPT to complete school assignments. 

But with generative AI continuing to evolve and grow in its capabilities, it rightfully begs the question: do assessments need to change in the AI era?

We sat down with Sarah-Eleni Zaferis (Atomi’s resident teacher and Product Enablement Specialist) to dive into how educators should be responding to the rise of AI and how to pivot assessment styles to better test for the application of knowledge. 

How AI enables continuous, real-time feedback

One of the best use cases for AI in the context of assessments is providing students with continuous, real-time feedback. 

Traditionally, a teacher might have 10 minutes per lesson (if they’re lucky) to check in on students individually to see how they’re tracking with curriculum content. In a classroom of up to 30 students, it’s not surprising that the majority of students can go weeks between gaining tailored, 1:1 feedback. 

With AI and tech, ongoing monitoring of students’ progress is possible—even at scale. From active recall quizzes to exam practice, AI and auto-marking technology ensure students are receiving timely formative feedback on where their strengths and weaknesses are.

Not only does this speed up the marking process (and remove lag time between submitting assignments and receiving feedback), but it empowers teachers to rethink what assessments look like, too. With AI, educators have a broader scope for what types of assessments they can run, such as facilitating online simulations for science experiments. 

Shifting from testing for understanding to applying knowledge 

But AI isn’t just helpful in the context of providing feedback. The rise of GenAI calls into question the modus operandi for assessing students, fast-tracking a shift that has already been taking place around structuring assessments. 

Many educators have been looking at how assessments can evolve to better reflect the skills needed in the 21st century (from critical thinking to problem-solving and collaboration).

No longer are educators focused on testing for understanding with hand-in projects and take-home essays. In the AI era, it’s about shifting to ensure students are capable of applying new knowledge.

It’s about testing for application of knowledge in unknown situations and combining threads of different pieces of knowledge… AI is speeding up that process.
- Sarah-Eleni Zaferis, Teacher and Product Enablement Specialist

The biggest benefit of this shift? While GenAI can be used by students to generate responses or synthesise knowledge, it isn’t able to effectively apply knowledge to new situations. By shifting to assessments that focus on applying knowledge, educators can productively reduce the risks of students using AI in unethical ways.

Practical tips for adapting assessments in the AI era

For educators who are ready to embrace the AI era, it’s time to pivot the way assessments are structured. 

In practical terms, here’s where Sarah-Eleni recommends teachers should start: 

• Look for quick wins: Remove the high stakes and start experimenting with the use of AI in non-assessment contexts to get used to the tools first. For example, you could use Chat GPT to produce practice material for students, including sample answers and marking criteria. 

There’s no risk involved in this scenario as it is non-assessable, but it gives teachers an opportunity to get familiar with AI and be creative with how it is implemented.
Sarah-Eleni Zaferis, Teacher and Product Enablement Specialist

• Design assessment that requires application of knowledge: Shift away from assessments that test for understanding (a high-risk case for the unethical use of AI) and focus on assessment types like problem-solving questions or even in-class debates that require students to applying knowledge in practical scenarios.
  
  
• Help students use AI appropriately in assessments: For tasks where students need to prepare notes or aids, support students to understand how to use AI to create these resources. Then, ensure the assessment is focused on the application of that knowledge to unknown situations in real time.
  
  
• Fast-tracking marking: With AI and tech, it’s possible to automate the feedback process for students and ensure continuous feedback is provided throughout the learning journey (rather than waiting until assessable moments). For teachers, this approach reduces the manual work involved in marking to free up time and resources that can be reallocated to preparing creative, innovative learning experiences in the classroom.

• Consistency of feedback: AI ensures that feedback is not just one touch point at the end of a unit. Instead, teachers are able to continually check in on progress and easily identify knowledge gaps as well as areas for improvement, ensuring students receive tailored revision resources to focus their learning moving forward.
  
  
• Gain whole-class insights: With tech platforms like Atomi, educators can easily track student performance at a class-wide level, quickly identifying trends in the data and tailoring class plans to suit students' levels of understanding. 

Looking ahead: the future of AI and assessments

So, what’s ahead in the world of AI and assessments? We’re only beginning to tap into the potential of tech and AI for educators and students, meaning there is plenty of room for innovation and growth. 

Enabling continuous assessment is one of the biggest trends for AI and assessments. With the help of AI-powered feedback and marking, students will be able to demonstrate growth across a whole year—rather than only offering one chance to demonstrate knowledge. 

At Atomi, we understand the impact of feedback and that true value comes from helping students understand why they received the mark they did, and how they can improve going forward. But we also understand that providing personalised feedback can be time-consuming for teachers. No matter how much feedback even the most dedicated teacher can provide, there's always a demand for more.

That’s why at Atomi, we use AI to provide feedback for short answer questions. Find out more about our latest AI-powered feedback product launch. 

‍