May 21, 2024
The rise of artificial intelligence (AI) has transformed many facets of our lives, including the way we learn and develop new skills. Scrum, a popular framework in the Agile world, has not been immune to this digital revolution. And with the advent of advanced AI engines, there's a growing question within the Agile student community: can AI be used to cheat in Scrum exams?
But more importantly, should it be?
A Brief History of AI and GPTs
AI's journey began in the mid-20th century, with initial research focusing on creating machines that could mimic human intelligence. Over the decades, AI has evolved dramatically, leading to the development of neural networks, machine learning, and more recently, Generative Pre-trained Transformers (GPTs).
The GPT series, developed by OpenAI, represents some of the most advanced AI capabilities to date. Starting with GPT-1 in 2018 and progressing to the highly sophisticated GPT-4, these models are capable of understanding and generating human-like text based on vast amounts of data. Other AI engines like Google’s BERT, Microsoft’s Turing, and IBM’s Watson have also made significant contributions to the field.
The Temptation of Using AI to Cheat
The capabilities of AI engines like GPT-4 are impressive. They can generate text, answer questions, and even simulate human conversation with remarkable accuracy. This has led to the temptation among some individuals to use these tools to pass online exams, including professional qualifications in Scrum. The theory being that you have AI open alongside your exam browser and type or read the questions you’re being asked into the engine to generate an immediate answer.
Scrum exams, such as those offered by Scrum.org and the Scrum Alliance, are designed to test a candidate’s understanding of Scrum principles, roles, and practices. Given the open-book nature of some of these exams and the ease of access to AI tools, it might seem like a straightforward path to success. Apart from the advanced qualifications, the exams are mostly multiple choice - ideal for throwing into an AI engine, you’d have thought.
Why Using AI to Cheat is a Bad Idea
Inaccuracy of AI Information: While AI has made tremendous strides, it is not infallible. AI-generated responses can sometimes be inaccurate or misleading. Quite often - possibly more often than we realise - AI responses contain incorrect information or “hallucinations”. After all, it’s using a vast array of sources to create its answers - not all of these sources will be accurate. So relying on AI to tell you which answer to tick to pass an exam could easily result in you selecting the wrong answers.
Shortcuts Undermine Learning: The primary purpose of taking a professional qualification is to learn and internalise the subject matter. Using AI to cheat shortcuts this learning process. Without truly understanding the principles of Scrum, a practitioner cannot effectively implement them in a work environment. This not only diminishes personal growth but also hampers the effectiveness of Scrum teams and projects.
Workplace Implications: Obtaining a Scrum certification without a genuine understanding of the framework can have serious repercussions in the workplace. Scrum Masters or Product Owners who lack foundational knowledge can derail projects, leading to inefficiencies, misunderstandings, and potential project failures. This undermines the credibility of the individual and can damage the organisation's trust in Scrum frameworks.
Ethical Considerations: Cheating, whether in a traditional exam setting or an online professional qualification, is unethical. It compromises the integrity of the certification process and devalues the hard work of those who have earned their qualifications honestly. Professional certifications are meant to be a testament to one's skills and knowledge; using AI to obtain them dishonestly is a breach of this trust.
A Real World Test Case
We didn’t write this blog just to encourage you to take courses, such as those that we offer (although we would very much recommend you do if you want to maximise your chances of success in your Scrum studies). We actually tested AI pretty thoroughly to come to our conclusions. We ran tests using mock PSM-1 exams and ChatGPT-4 and found that the hit rate for correct answers was 50-50.
We went a step further and created a bespoke GPT on one of our laptops that we trained solely on legally purchased Scrum books and officially published blog posts - the kind of books and articles that you’d likely read when studying for your certifications. This meant that our engine was only using official materials and excluded the general morass of Agile internet content, such as blogs, opinion pieces, news articles and forums on which AI often relies to draw its conclusions when answering your questions. The hit rate improved, but not by as much as you’d expect: let’s just say that you’d be very close to the pass/fail line if our testing is indicative of reality.
A Better Approach to AI in Learning
Instead of using AI to cheat, individuals should leverage AI as a learning aid. AI can be a powerful tool for studying and understanding complex topics. For instance, AI-driven platforms can provide personalised learning experiences, identify knowledge gaps, and offer explanations and resources to deepen understanding.
Study Assistance: AI can help break down complex Scrum concepts into more manageable pieces, providing explanations and examples that enhance comprehension. Using AI for practice tests and quizzes can help reinforce knowledge and identify areas that need more focus. But per our test results above, be sure to do your own fact checking on the outputs from the AI engine.
Interactive Learning: AI-powered chatbots can simulate real-world Scrum scenarios, allowing learners to apply their knowledge in a practical setting. This interactive approach can enhance learning outcomes and better prepare individuals for both exams and real-world application.
Ethical AI Usage: Embracing AI ethically means using it to augment one’s study habits rather than replace them. This approach ensures that individuals gain the full benefit of the learning process and are genuinely prepared for their roles in Scrum teams. By doing your own legitimate studying, you’ll quickly learn how to use AI to augment your study and work: you’ll develop an ability to distinguish between valuable information and AI hallucinations in the answers it gives you.
Conclusion
While AI tools like GPT-4 offer immense capabilities, using them to cheat on Scrum exams is neither wise nor ethical. The true value of a professional qualification lies in the knowledge and skills acquired during the learning process. By leveraging AI as a study aid rather than a shortcut, individuals can ensure they are fully prepared to implement Scrum effectively in their work environments, thereby upholding the integrity of the certification and the principles of Agile frameworks. Talk to us today if you’re interested in taking a Scrum course.
