The Next Generation of Item Development: Can AI item generation benefit your program?
A blog by Nikki Eatchel, Chief Assessment Officer, Prometric
The Issue of Security
While technological advancements are increasingly met with new and emerging cheating tactics, this is just one of the many security threats that often plague testing organizations.
These threats present more than just a danger to your test items; they also pose a serious threat to the very reputation of your organization’s credentials. The financial implications of security breaches are also significant; this year alone, the average cost of a security breach increased 2.6% from $4.24 million in 2021 to $4.35 million in 2022.
However, when Artificial Intelligence (AI) is leveraged, the average cost of a data breach can be reduced by up to $3.05 million. With evolving cheating tactics such as item harvesting, proxy testing, and more advanced technology-driven measures, such as wireless Bluetooth earbuds being used to communicate with a third-party during the exam, the list of threats your organization may face is ever-growing.
To protect your organization’s reputation, and to ensure the success of your assessments, it’s vital to have the proper technology in place to combat these emerging threats throughout every stage of the exam development and delivery lifecycle – starting with exam development. Today, AI tools and techniques, such as computer-adaptive testing or Automated Item Generation (AIG), have long been leveraged within exam development to provide an additional layer of security for exam items, but as the testing landscape evolves, these tools begin to fall short on providing the layer of defense they once offered, in addition to providing their own challenges.
How many items do we need in our item bank to reduce exposure and maintain exam integrity? How often should items be refreshed to reduce exposure and increase score integrity? How many Subject Matter Experts (SMEs) do we need to recruit to properly cover our test domains? While these are some of the most common questions many testing organizations begin asking early in the exam development process, it’s also important to consider the additional challenges these questions present, such as SME burnout, turnover, human bias, novelty of content, lack of agility in changing exam topics, etc. It’s no wonder exam development can quickly drain your time, money, and labor resources… That is, until now. What if you could increase productivity, creativity, and the speed of item authoring, all while streamlining SME time to develop content, eliminating the additional time and costs of generating questions from scratch? While this may sound ideal, and too good to be true, the one outstanding question that remains is how does this work?
The Emergence of AI
The next generation of test development utilizes AI-assisted item generation to reduce the time needed to build and bank a range of quality items. The software works with your SMEs (not as a replacement of your SMEs) to continually enhance the software’s understanding of your program needs, and to subsequently generate higher quality items the more it learns about your exam content. By implementing this type of groundbreaking AI, your organization can streamline the item development process, creating a higher volume of quality exam items at up to 10X the speed. In turn, this allows your SMEs more time to spend reviewing and enhancing AI-generated items, all while refining the AI model to deliver better results. No longer do SMEs need to spend the majority of their time digesting exam domains and creating items from scratch. While this may sound great, the looming question from above still remains: how does it work? The suite of AI tools from Finetune allows your program to generate quality content without relying on templates, clones, or variants, in order to provide unique items, based on the domains required for your exam. Rapidly producing complex and authentic scenarios and situations could previously only be acquired by employing multiple SMEs to review very detailed material in the required subject area before draft ideas could even be produced. Finetune’s AI model is focused on providing secure and efficient item development, without sacrificing quality.
- Security: Reduce the possibility of breaches while simultaneously providing SMEs with quality test items produced at a rapid pace. Build out a greater volume of AI-generated items for your assessments while building an additional layer of security through reduced item exposure, ensuring the longevity of your exam forms and preserving the prestige of your credentials.
- Efficiency: Produce items up to 10 times faster than a conventional item writing approach. In addition, the AI model also takes less time to establish, while eliminating the costs of copyright permissions, significantly reducing your organization’s time and investment costs, such as those incurred by running frequent writer workshops.
- Quality: The AI model increases creativity and variation in items, without the use of templates, eliminating clones and variants (and expanding the breadth of depth of your item banks). Leverage novel scenarios that would typically only be achieved through extensive SME collaboration, all while maintaining a stringent focus on quality. Finetune’s model is built to provide you the flexibility your program needs, learning from your SMEs, and integrating their competency, taxonomy, and cognitive complexity with each use.
The significant investment organizations make in their credentials is measured not just financially, but in time, effort, perception, and reputation. While the cost of generating items is a substantial portion of running and maintaining a successful licensing or credentialing program, consider the additional costs incurred in each of these areas by security breaches, and that price far exceeds the average of $4.35 million in 2022 alone. By leveraging Finetune’s AI-assisted item generation, your organization is safeguarding more than just your exam assets, it is ensuring the ongoing vitality and relevance of your credential.
So, how do you determine if Finetune’s AI item generation can benefit your program? While this innovative AI approach may seem intimidating to grasp, the implementation and use of the tool is anything but intimidating. Not only can Finetune’s AI model strengthen the safety of your exam with fewer physical touchpoints, but also enhance exam development with creative item generation and an extensive item pool, extending exam life through reduced item exposure. By leveraging Finetune’s AI-assisted item generation, safeguard more than just exam assets, and ensure the ongoing vitality, relevance, and security of your credentials.