Earlier this year, we noticed something strange about the job applications coming into The Learning Agency: they all felt remarkably alike, as if the applicants had collaborated on a template cover letter and resume. Our hiring team found it difficult to distinguish one candidate from another.
They suspect the applicants were relying heavily on ChatGPT, or one of the other popular AI platforms like Claude or Gemini.
This is not unlike the dilemma facing educators and schools, where students are increasingly using AI to assist them in completing their assignments. How’s a teacher to know if students are actually learning or are just passing off AI-generated work as their own?
In both circumstances, the solution is not to lash out at AI or to institute draconian rules limiting the technology’s usage. Rather, the most reasonable and realistic path forward is for decision-makers, like schools and employers, to rethink their operations to account for AI use.
One of the best and most effective ways to do that is to incorporate live assessments into their evaluation processes. Live assessments such as oral defenses, real-time project planning, on-the-spot problem-solving, or collaborative group work allow educators and employers to observe how individuals think, adapt, and communicate in the moment.
At The Learning Agency, this means asking job candidates to perform a final screening task where they have to verbally respond to a question related to an answer they had previously submitted. Candidates are given one minute to read the prompt, one minute to prepare, and one minute to deliver their response on video.
Such an approach requires candidates to be prepared and comfortable enough to deliver concise, thoughtful responses under time constraints, and gives evaluators insight into how they respond to different scenarios without outside help or support. As a company, our hiring managers have found these tasks helpful in highlighting each candidate’s unique qualities.
Similar opportunities await in education. A 2024 report from Inside Higher Ed found that nearly half (47%) of the 2,000 students polled said that generative AI has made it much easier to cheat. As concerns over AI-enabled cheating and plagiarism rise, teachers should shift their focus toward real-time, live assessments that measure genuine student understanding.
One of the best and most effective ways to do that is to incorporate live assessments into their evaluation processes. Live assessments such as oral defenses, real-time project planning, on-the-spot problem-solving, or collaborative group work allow educators and employers to observe how individuals think, adapt, and communicate in the moment.
Blurring The Lines Of Academic Integrity
As AI tools become more accessible, the boundary between acceptable support and academic dishonesty is increasingly unclear. Is it cheating to use AI to brainstorm ideas or edit a draft? When does assistance become misrepresentation? What qualifies as original student work in the age of AI?
Rather than banning these tools, educators should rethink how they assess learning. If the goal is to determine what a student truly knows and understands, teachers should look to include assessments that are insulated from AI misuse, and that often means doing it live.
Economist David Deming sees an opportunity to change assessments for the better in the age of AI. He offers an important insight: the use of AI to complete assignments may indicate that the tasks teachers have been assigning are too shallow. He argues that AI will push educators to shift toward deeper forms of learning that require students to embody knowledge and apply it meaningfully.
“You might imagine in a classroom, almost returning to an older tradition of oral education,” he said. “We focus more on, can you make a presentation? That’s AI proof because…you have to actually be able to do it in the moment and defend your position.”
Economist David Deming sees an opportunity to change assessments for the better in the age of AI. He offers an important insight: the use of AI to complete assignments may indicate that the tasks teachers have been assigning are too shallow. He argues that AI will push educators to shift toward deeper forms of learning that require students to embody knowledge and apply it meaningfully.
This perspective is also echoed by Carie Cardamone, Associate Director for STEM & Professional Schools at Tufts University, who notes in an article for Teaching@Tufts that AI exposes weaknesses in traditional assessment methods that rely on independent, out-of-class work. Cardamone highlights the need to re-envision assessments, integrating in-class activities that demonstrate knowledge with traditional assignments, which will make it more challenging for students to rely solely on AI.
What both Deming and Cardamone are describing is a form of live assessment.
Traditionally, student assessment comes in two forms: formative and summative. Formative assessments measure a student’s progress over time, providing ongoing feedback that helps identify strengths and weaknesses through activities like journal responses, class discussions, or exit tickets. Summative assessments, on the other hand, evaluate learning at the end of an instructional period, often through midterm exams, final projects, or standardized tests.
As AI permeates teaching and learning, the next step forward is live formative assessments.
Live Formative Assessments
Live formative assessments require teachers to monitor student progress in real time, with the potential to provide immediate feedback. They go beyond take-home tests, projects, and essays, which are vulnerable to AI misuse due to the lack of supervision and difficulty verifying authorship. Live assessments offer deeper insight into student understanding and allow educators to authentically assess learning.
For example, instead of assigning a reading and then assigning students a four-page essay on the piece, students could be asked to develop their essay in class while engaging in discussions about key symbolism and receiving peer feedback.
The advantage here is that teachers can directly observe how students build arguments, solve problems, and communicate ideas. This not only supports academic integrity but also encourages students to engage meaningfully with the content.
It’s important to recognize that live assessments may not work for every classroom. Students who are English learners, neurodivergent, or who experience performance anxiety may struggle with real-time tasks, especially those that involve being in front of their peers. However, live assessments can be designed in ways to accommodate a learner’s needs, such as giving students more time with tasks, having students draw from their lived experiences, or working in small groups.
Live assessments may not work for every classroom. Students who are English learners, neurodivergent, or who experience performance anxiety may struggle with real-time tasks, especially those that involve being in front of their peers. However, live assessments can be designed in ways to accommodate a learner's needs.
Preparing Students For An AI World
Generative AI has changed the educational landscape, both as a challenge and an opportunity. As I’ve seen in hiring here at The Learning Agency, live assessments can help uncover the human voice behind the resume. The same is true in education. By investing in live formative assessments, we can help protect academic integrity, promote deeper learning, and prepare students for a future where AI is a tool, not a shortcut.
This column first appeared in Forbes.
