Protecting assessment integrity in the age of generative AI

Education is a vital pillar in shaping the next generation and their outlook on life. And in today’s world, assessments act as the key metric for determining educational achievement. These assessments serve as a standardized way to gauge how effectively an individual has absorbed and applied their learning.

Since candidates now have access to sophisticated AI tools capable of generating complex solutions, our traditional assessment security measures must evolve to keep pace.

The rise of artificial intelligence is rapidly transforming the assessment landscape, making it simpler than ever for candidates to secure instant answers. Candidates can call in real-time answers to the most intricate questions in just a few clicks from ChatGPT. From simply pasting questions into ChatGPT to leveraging voice-assisted applications like Cluely, the challenge of maintaining exam integrity has never been more pressing.

The era of AI vs anti-AI is here, and we have the ultimate solution against the use of AI cheating tools – MSB. Mercer Secure Browser is specifically designed to shut down these hidden threats, ensuring the validity of your online assessments.

There are broadly two categories of AI applications that require blocking to ensure integrity in online assessments.

Category A: These are applications that require text input to generate answers. ChatGPT is a prime example here. The candidate is required to paste the question in the context area to get the correct answers. Access to these applications during assessment can be restricted using MSB by implementing an unbreachable lockdown mode. If a candidate cannot move out of the test window, the question does not reach the ChatGPT window.

Category B: These are applications that can take voice input, such as Cluely, or can be invoked by shortcut commands to capture questions on screen. Tools such as InterviewCoder and LeetCode Wizard offer seamless real-time help during live coding assessments, remaining invisible during screen share while generating code, explanations, and even time/space complexity. The classic lockdown method cannot effectively combat these modern AI threats, as legacy lockdown solutions are blindsided in this context. However, MSB integrates multiple defense overlays to ensure a secure, verifiable exam environment.

1. Overlay protection: MSB has sustainable control over screen permissions and enforces strict checks, which allow us to intercept and block overlays before they can compromise the exam environment. Proactive prevention, not reactive detection. MSB proactively prevents cheating tools from being launched in the first place.

2. Universal cheating software database: Over the past decade, MSB has maintained and continues to expand a global library of cheating applications. With us, all mainstream cheating applications are already off the table.

3. Active process monitoring: MSB employs a lightweight orchestration engine that ensures no background running process interferes when the test is in progress. If an unknown process attempts to inject itself into the exam space, we detect and block it immediately, rendering the process inert for the duration of the test.

• MSB is the most trusted, tested, and future-ready secure browser available. What sets us apart is that the browser is not built on assumptions, but rather on real-world experiences and feedback from our clients.
• MSB is available on both Windows and Mac and is designed for plug-and-play integration with any assessment platform.

With MSB, you’re not just preventing cheating; you’re redefining the integrity of online assessments.