Guest post by Melissa DeWees, Director of AI Ethics and Compliance at Proctorio
Artificial Intelligence (AI) is no longer the villain of sci-fi stories; it has now become an integral part of everyday life, quietly emerging as the unsung hero of our generation. Achieving this transformation wasn’t easy. It took a lot of time and exposure to AI’s innovative capabilities for people to warm up to it. AI has played a key role in revolutionizing industries by fostering innovation, improving efficiency, and expanding the potential of various technologies.
Despite their advanced capabilities, AI systems function within the limits set by their human creators, with their actions and abilities defined by the algorithms and data they are trained on. When used ethically and transparently, AI can profoundly impact various aspects of life. In transportation, for example, it optimizes traffic flow, improves navigation, and enhances road safety. In areas such as education and healthcare, AI is streamlining processes and providing solutions that were once considered theoretical. In healthcare, machine learning aids early disease detection, improves diagnostic accuracy, and supports personalized treatment plans. Education also benefits from AI, with novel applications in writing assistance, test design, question generation, automated grading, plagiarism detection, and even online proctoring.
From transportation to education and beyond, AI has the potential to enhance everything it touches. However, it’s important to remember that while AI offers significant benefits, it cannot—and is not meant to—replace human oversight. AI’s primary purpose is to make our lives easier.
AI functions as a tool that reflects the intentions of its creators, and this principle is evident in how Proctorio leverages AI. Our primary use of AI is in online exam proctoring, which monitors exam activity, providing credibility to online education. Based on the settings chosen by the exam administrator, Proctorio conducts realtime integrity monitoring. It can identify if more than one person is in the room, detect additional voices, or notice if the test-taker leaves. It can also detect exam collusion, use of virtual machines, and restrict VPN activity. Browser lockdown features prevent access to additional tabs or unpermitted resources. The AI alerts administrators to any unusual activity, ensuring a secure testing environment.
Our facial detection algorithms, trained using data with consent, utilize human annotators to label attributes and create balanced, diverse datasets. We’re actively working to ethically source AI databases to reduce biases (a major concern in artificial intelligence).
Biases in AI can arise from algorithmic or data flaws, often caused by skewed or incomplete datasets. Societal biases, stemming from human assumptions and blind spots, can also affect outcomes. To combat these issues, it’s essential to address biases during the AI training process. At Proctorio, we had an Independent third-party evaluate our algorithm that showed no statistically significant bias toward any group. We are even committed to have our systems audited. However, we aim to lead in this space, so we continually improve our solutions and products.
Proctorio’s use of AI sets us apart from other online proctoring companies. Our system is highly scalable, requiring no human proctors while handling large volumes of exams simultaneously. This makes it easier to manage peak testing times and allows global accessibility. AI-automated proctoring not only reduces costs for test takers and institutions but also minimizes the potential for human bias by applying consistent, objective criteria based on the exam administrator’s settings. Additionally, real-time analytics are generated for immediate administrator review.
When developed and deployed responsibly, AI has the potential to transform how we interact with the world daily, making tasks easier and freeing our time for the things we find important. While we embrace AI’s technological advances, we must also remain mindful about its ethical implications. The future of AI is still unfolding, and it’s our responsibility to ensure it remains a positive change.
