Modern learning technologies have opened new doors for learners, but they have also empowered the rise of sophisticated ai cheating tools powered by generative AI (Gen AI). In 2024, a global survey found that 56 % of students admitted to using AI tools such as ChatGPT, or Claude during exams. These tools can generate answers in real time, even slip past traditional plagiarism checks. With fraud risks predicted to climb from 33 % in 2025 to 81 % by 2028, traditional proctoring simply isn’t enough.
To address this crisis, Talview has secured a U.S. patent for Alvy, the world’s first agentic AI proctoring software. Unlike rule-based monitors, an agentic AI behaves as an autonomous agent that perceives, reasons and acts frequently. Alvy’s design combines the power of Agentic AI with a seven-layer security framework, creating a robust, adaptive shield against modern cheating methods.
Why Alvy as Agentic AI Proctoring?
Conventional proctoring systems rely on fixed or static rules, flagging eye movement or background noise and passing them to a human reviewer. These passive systems struggle to keep pace with evolving AI cheating fraud. By contrast, an agentic AI proctoring technology acts like an intelligent autonomous guardian. It learns from each assessment, adapts to new Ai cheating tactics and makes context aware decisions in real time.
Alvy embodies this paradigm shift. It employs six specialized features for identity verification, live monitoring, real time candidate assistance, research and ai cheating pattern detections. This modular architecture allows the system to operate in both Copilot and Autopilot modes. Copilot assists human proctors, doubling their coverage and cutting review times by up to 75 %, while Autopilot delivers fully automated monitoring with a 40 % reduction in false alerts
Decoding the Talview Seven Layer Security Framework
To understand what makes this patented technology different, it helps to look at its layered defense:
- Identity verification – Real-time face recognition and photo ID matching prevent impersonation and deepfakes.
- Behavior monitoring – AI video analysis detects offscreen glances, gestures, and whispering.
- Environment monitoring – Secondary cameras enable 360° room scans to spot hidden people or devices.
- Device security – Secure browser blocks internet searches, VMs, ai cheating tools, multiple monitors and screen sharing.
- Assessment feed monitoring – Real-time checks flag AI-generated answers and irregular navigation.
- Intelligence layer – Cross-session analytics detect collusion, identity fraud, and repeat cheating.
- Web monitoring – Scans the internet for leaked questions and shared answers before, during, and after exams.
By layering these defenses system, Alvy proactively detects, prevents and deters both simple impersonation and advanced AI cheating fraud.
What’s Inside the Patent?
Talview’s Alvy Patent (Patent No. 12,361,115 B1) covers the architecture and functioning of Alvy’s agentic AI engine. Unlike traditional proctoring systems that simply flag predefined violations, Alvy operates as a closed loop agent that observes, interprets, decides and acts autonomously.
The patent system incorporates multiple subagents that gather data from video, audio, screen activity, identity checks, browser interactions and environmental signals. These inputs feed an AI-driven risk scoring engine, which uses contextual awareness and cross-session analytics to detect collusion, identity fraud and pattern-based cheating.
When a threshold is met, Alvy can intervene in real time, either by assisting the candidate, escalating to a human proctor or autonomously blocking suspicious activity. The patent specifically protects this agentic control loop, ensuring that only Talview can lawfully deploy such a combination of continuous monitoring, adaptive decision making and autonomous action. In practical terms, this means Alvy isn’t just an algorithm that watches it’s a network of intelligent agents that perceive, decide and act to preserve exam integrity, and its patented status provides legal clarity for organizations adopting it.
What’s Inside for you?
For universities, certification bodies and enterprises, the patented agentic AI delivers multiple benefits:
- Legal and competitive advantage – Because this technology is patented, organization adopting Alvy can be confident they are using a legally protected innovation that differentiates them from competitors. No other proctoring vendor can replicate Alvy’s agentic architecture.
- Compliance and transparency – The seven layer framework cover identity verification, device security, behavior monitoring, environment scanning, assessment feed analysis, intelligence-based pattern detection and web monitoring. Each layer produces auditable data, helping institutions meet regulatory requirements and demonstrate fairness and transparency.
- Improved trust and candidate experience – Because Alvy provides Realtime assistance and reduces false positives by 40 %, it both enhances integrity and improves candidate satisfaction. The system’s ability to detect 99 % of events that human proctors see assures stakeholders that assessments are both accurate and unbiased.
- Operational efficiency – With Copilot mode doubling proctor coverage and cutting review time by up to 75 %, organizations can scale exams without proportional increases in staffing. Autopilot mode handles medium stakes tests independently, reducing human workload and freeing proctors to handle exceptions.
How the Open edX LMS Benefits from Alvy?
The Open edX LMS, Talview’s core partner and one of the world’s leading open-source learning platforms, empowers millions of learners and educators globally. With the integration of Alvy’s patented agentic AI proctoring, the Open edX LMS can now deliver secure, scalable, and user-friendly proctoring directly within its platform.
For universities, training providers, and enterprises using the Open edX platform, this means:
- Seamless Integration – Exams remain within the familiar Open edX environment.
- Unmatched Security – The seven-layer framework blocks both traditional and Gen AI cheating tools.
- Global Scalability – Autopilot and Copilot modes allow mass assessments without overburdening staff.
- Transparency & Trust – Audit-ready reports enhance credibility for online credentials.
This technology collaboration reinforces the Open edX community’s mission to make education accessible while ensuring academic integrity at scale.
“The Open edX Project is committed to advancing accessible and credible learning experiences.”
– Said Eden Huthmacher, Open edX Community Manager
Future of Trust-Driven Proctoring Starts Now
Alvy’s patent marks more than an innovation milestone, it redefines what AI in assessments, exams should embody: intelligence, autonomy, ethics, and transparency. As Ai cheating technologies rapidly evolve, only equally advanced, agentic systems can deliver the scale, speed, and sophistication needed to maintain exam integrity.
“This patent reaffirms Talview’s commitment to building secure, fair, and candidate-friendly assessments.”
– Said Sanjoe Jose, CEO of Talview
By combining Agentic AI with our deep domain expertise, we’re helping organizations using the Open edX platform deliver assessments that inspire trust at scale.
Whether you’re managing certification exams, academic exams, or enterprise hiring assessments, Alvy provides the assurance that every evaluation is protected by the world’s first patented Agentic AI.
- Alvy Patent Details to explore the technical innovation behind our breakthrough
- Experience Alvy’s Agentic AI and see the difference intelligent proctoring makes
