Trust & ethics of AI use

We believe artificial intelligence can significantly improve how people learn, work, collaborate and access digital services. At the same time, we recognise that AI introduces new responsibilities and risks, particularly when it is used in education, assessment, examination invigilation, identity verification, productivity and other environments where technology can influence important decisions.

This Trust & Ethics of AI Use statement describes the principles that guide how Ann-Connect develops, evaluates, deploys and uses artificial intelligence across our products and services. It is intended to provide customers, institutions, users, educators, administrators and other stakeholders with greater transparency about our approach to responsible AI.

Our objective is not simply to build powerful AI systems. We aim to build AI systems that are useful, secure, understandable, appropriately governed and designed with people in mind.

Artificial intelligence is increasingly becoming part of the digital infrastructure organisations rely on. At Ann-Connect, we see AI as a tool that should augment human capability rather than unnecessarily replace human judgement.

Our AI capabilities may be used to support activities such as learning, productivity, content assistance, analysis, workflow automation, examination processes, invigilation and other platform functions. The specific capabilities available to a customer may depend on the Ann-Connect product, configuration and applicable contractual arrangements.

We recognise that AI systems are not infallible. They can produce inaccurate, incomplete, inconsistent or inappropriate outputs. They may also perform differently across different contexts, languages, environments or populations.

For this reason, we design our AI approach around appropriate human oversight, transparency, privacy, security, fairness and accountability.

We believe that AI should support people rather than remove appropriate human responsibility.

Where an AI system produces an output that may have meaningful consequences for an individual, organisations should consider the appropriate level of human review before taking consequential action.

This principle is particularly important in educational environments. An AI-generated alert, classification, recommendation or observation should not automatically be interpreted as definitive proof of academic misconduct or other wrongdoing.

Where Ann-Connect's technology identifies behaviour or activity that may require attention, the appropriate institutional processes and human judgement should remain important parts of the overall decision-making process.

AI can help identify information that a human may need to review. It should not be treated as a substitute for reasonable investigation, institutional policy or professional judgement.

We believe users should have a reasonable understanding of when artificial intelligence is being used and what role it plays within a product or workflow.

Where appropriate, Ann-Connect aims to provide information about AI-powered functionality through product interfaces, documentation, contractual materials or other relevant communications.

We do not believe that simply describing a product as “AI-powered” provides sufficient transparency. Responsible use also requires understanding the purpose and limitations of the technology.

Our objective is therefore to communicate AI functionality in a way that is understandable to the people and organisations using our platform.

AI systems can make mistakes.

An AI-generated response, classification, recommendation, detection or prediction may not always be correct. The reliability of an AI system can depend on factors including the quality of the available information, system configuration, environmental conditions, model limitations and the specific context in which the system is used.

Ann-Connect therefore does not represent that AI-generated outputs are inherently accurate, complete or suitable for every purpose.

Users and organisations remain responsible for applying appropriate judgement when relying on AI-generated information, particularly where the outcome could materially affect an individual.

We continuously seek opportunities to evaluate and improve the performance of our AI capabilities while recognising that improvement does not eliminate the possibility of errors.

Assessment and examination environments require particular care because technology can potentially affect a student's academic experience and outcomes.

Where AI is incorporated into Ann-Connect's assessment or invigilation capabilities, the purpose is to assist institutions in identifying information or activity that may warrant further review.

AI-based indicators should be interpreted within the appropriate context. A technical anomaly, unusual behaviour or system-generated flag does not necessarily establish that a student has acted improperly.

Factors such as connectivity interruptions, device limitations, environmental conditions, accessibility requirements and other legitimate circumstances may affect what an AI system observes or reports.

Institutions using Ann-Connect remain responsible for establishing and applying their own academic integrity policies, examination rules and review procedures.

Ann-Connect's technology is designed to support those processes rather than independently determine a student's guilt, innocence, academic performance or disciplinary outcome.

We recognise that AI systems can reflect limitations or biases present in their training data, underlying models, system design or operating environment.

This can create differences in system performance across different users, environments, languages, devices or other circumstances.

We therefore consider fairness and potential bias as important considerations when developing and evaluating AI-enabled functionality.

Where appropriate, we seek to identify potential sources of bias, evaluate system behaviour and improve our technology when issues are identified.

We also recognise that fairness is not a one-time technical exercise. AI systems evolve, models change and new use cases introduce new risks. Responsible AI therefore requires ongoing monitoring and review.

Privacy is a fundamental consideration when developing AI capabilities.

Some AI-enabled features may process information provided by users or generated through their interaction with Ann-Connect. Depending on the product and configuration, this information may include account information, educational information, user-generated content, examination information, communications, device information or other data necessary to provide the relevant service.

We seek to limit the collection and use of information to what is reasonably necessary for the applicable purpose and to apply appropriate security safeguards.

Our approach to personal information is also governed by our privacy and data protection commitments.

For information about how Ann-Connect collects, uses, stores and protects personal information, users should refer to our Privacy Policy and other applicable data protection documentation.

Where AI systems involve data used for evaluation, improvement, testing or development, Ann-Connect considers the applicable legal, contractual, privacy and security requirements governing that data.

We do not believe that access to customer information automatically means that such information should be freely reused for unrelated purposes.

Where customer data may be used for a particular AI-related purpose, the applicable contractual terms, product configuration and privacy requirements will determine how that data may be handled.

Customers with specific requirements regarding the use of their data should discuss those requirements with Ann-Connect before deploying relevant AI functionality.

AI systems introduce security considerations in addition to those associated with conventional software.

These may include risks relating to unauthorised access, manipulation of inputs, inappropriate outputs, data exposure, model behaviour and misuse of AI functionality.

We incorporate security considerations throughout the development and operation of our AI capabilities and seek to apply appropriate technical and organisational safeguards.

AI functionality is also subject to the broader security architecture and controls applicable to the Ann-Connect platform.

Our security practices may evolve as new AI-related threats and industry practices emerge.

Generative AI can produce text, summaries, recommendations, explanations, code, images and other forms of content. These capabilities can improve productivity and provide valuable assistance, but generated content may contain inaccuracies or omissions.

Users should review AI-generated content before relying on it in situations where accuracy, completeness or professional judgement is important.

AI-generated content should not automatically be assumed to represent the views, policies or official position of Ann-Connect, a customer's organisation or any other person unless explicitly confirmed.

Organisations using generative AI should also establish appropriate internal policies regarding acceptable use, confidentiality, academic integrity and information sharing.

Technology should strengthen academic integrity rather than undermine it.

Ann-Connect's assessment and invigilation capabilities are designed to support institutions in creating controlled and trustworthy assessment environments. However, technology alone cannot establish academic integrity.

Academic integrity ultimately depends on a combination of institutional policies, assessment design, educator judgement, student responsibility and appropriate technological controls.

AI-generated observations should therefore be considered as part of a broader academic integrity process rather than as an automatic disciplinary mechanism.

Institutions remain responsible for determining what constitutes academic misconduct and how potential violations should be investigated and addressed.

Responsible AI should be accessible to the people it is intended to serve.

We recognise that users may have different abilities, devices, languages, connectivity conditions and technological environments.

When developing AI-enabled features, we consider accessibility, usability and inclusion as part of the broader product development process.

We also recognise that certain AI systems may perform differently depending on environmental or user-specific factors. Where such limitations are identified, we seek to improve the experience and provide appropriate alternatives or human processes where necessary.

Connectivity resilience is a core part of Ann-Connect's technology philosophy.

We recognise that AI-powered functionality may depend on computing resources, network availability, local device capabilities or other technical infrastructure. Where applicable, Ann-Connect designs products to support resilient operation in environments where connectivity may be unreliable.

However, offline functionality does not necessarily mean that every AI capability can operate completely offline. The availability of particular AI features may depend on the product, technical architecture and configuration being used.

We aim to communicate these limitations clearly so that customers can make informed decisions about how AI functionality should be deployed.

Certain Ann-Connect capabilities may rely on third-party technologies, infrastructure, models or services.

Where third-party AI technologies are used, their operation may be subject to additional technical, security, privacy and contractual considerations.

Ann-Connect evaluates third-party technologies as part of its technology and risk-management processes and seeks to use providers that are appropriate for the intended application.

The specific third-party services applicable to a customer's environment may depend on the products and configuration selected.

Responsible AI requires governance beyond the engineering team.

As our AI capabilities develop, we intend to establish and maintain appropriate processes for evaluating AI-related risks, reviewing new AI functionality and responding to identified issues.

These processes may include technical testing, security reviews, privacy assessments, performance evaluation, human oversight requirements and documentation.

Our AI governance practices will continue to evolve alongside changes in technology, regulation and industry standards.

We do not consider responsible AI to be a one-time certification or completed project.

AI technologies evolve rapidly, and systems that perform appropriately in one environment may require additional evaluation as they are expanded into new use cases.

We therefore seek to monitor the performance and behaviour of our AI-enabled functionality and use identified issues to inform future improvements.

Where appropriate, we may update models, system configurations, safeguards, documentation or user guidance to address emerging risks and improve reliability.

Responsible AI is a shared responsibility.

Customers using Ann-Connect are responsible for configuring AI-enabled functionality appropriately for their environment and for establishing suitable internal policies and procedures.

Customers should consider factors such as the purpose of the AI system, the sensitivity of information being processed, the potential consequences of AI-generated outputs and the level of human review required.

Customers should not use AI-generated outputs as the sole basis for consequential decisions where such use would be inappropriate under applicable law, institutional policy or contractual requirements.

We believe innovation and responsibility should develop together.

Our goal is not to avoid the opportunities presented by artificial intelligence, but to build and deploy AI in a way that recognises both its potential and its limitations.

As Ann-Connect grows, we will continue to refine our AI governance practices, improve transparency and strengthen the safeguards surrounding our AI-enabled products.

We expect this document to evolve as our technology, products, regulatory obligations and understanding of responsible AI develop.

We may update this Trust & Ethics of AI Use statement from time to time to reflect changes in our products, AI capabilities, technology, legal requirements or responsible AI practices.

When material changes are made, we may update the “Last Updated” date at the beginning of this page and provide additional notice where appropriate.

We encourage customers and users to review this page periodically to understand our current approach to responsible AI.

We welcome questions, concerns and feedback regarding how artificial intelligence is developed and used within Ann-Connect.

If you have questions about a particular AI feature, its purpose, the information it processes or how AI-generated outputs should be interpreted, please contact our team.

We believe AI should be powerful enough to help, transparent enough to understand, and responsible enough to trust.