Introduction
In the rapidly evolving technology sector, the integration of artificial intelligence (AI) has the potential to revolutionise technology platforms, streamline administrative processes, and enhance overall operational efficiency. Convertr recognises the importance of adopting AI technologies while ensuring their ethical and responsible use.
Convertr is committed to ensuring we use AI in a secure and responsible way, respecting privacy, confidentiality and third-party rights. This includes any AI used by third parties on our behalf. We use the term "artificial intelligence" to describe machine-based systems that infer solutions to set tasks and have a degree of autonomy. However, the scope of this policy includes any machine-driven insights resulting from the use of systems and techniques within the field of data science.
Purpose
The purpose of this policy is to establish clear guidelines for the development, implementation, and monitoring of AI systems to protect personal data, uphold ethical standards and mitigate potential risks. AI technologies must be used in a manner that aligns with legal requirements and maintains the trust and confidence of our stakeholders.
This policy outlines key principles and procedures that must be adhered to when utilising AI technologies within Convertr. It addresses critical aspects such as data privacy, algorithm transparency, accountability and ongoing monitoring of AI systems. As AI continues to evolve, this policy will be updated periodically to reflect advancements in technology, regulatory requirements and best practices.
AI Guiding Principles
1. Responsible Experimentation
We encourage responsible experimentation with AI tools to discover innovative solutions that advance our mission. Experiments should be conducted with appropriate safeguards and oversight. All AI-related projects must be reported to management and align with our organisation’s values and strategic goals.
2. Preventing Unfair Bias
Promoting fairness and non-discrimination is at the core of our business philosophy. We take proactive steps to prevent the creation or reinforcement of unfair bias by employing robust testing, extensive review and automated bias detection systems. This helps to ensure equitable outcomes across diverse populations.
3. Data Protection and Privacy
We comply with all relevant data protection legislation, including the UK & EU GDPR and the US CCPA. We ensure transparency in our use of personally identifiable information (PII) in AI, obtaining explicit consent where required. Sensitive or confidential data will not be input into third-party AI systems without appropriate safeguards.
In terms of AI data use specifics:
- No PII is included in AI training datasets.
- Each AI model is trained independently on a per customer basis to ensure no data is shared between customers.
- Data undergoes rigorous preprocessing, including encoding and standardisation, to maintain integrity and prevent exposure of raw data.
- Models do not store training datasets after processing to ensure customer data is not embedded or outputted.
4. Transparency and Accountability
We are committed to transparency in AI usage. AI-generated content or AI use in client-facing applications will be clearly labelled. We will maintain audit trails of AI usage and establish clear lines of accountability, including nominating senior responsible owners for AI projects.
5. Human Oversight and Editorial Control
Human accountability is fundamental to the development and use of AI. While AI may assist in decision-making, it will not replace human oversight. AI-generated content will be reviewed and edited by staff to ensure accuracy, fairness and compliance with institutional values before publication or implementation.
6. Monitoring Bias and Fairness
AI systems can perpetuate or amplify biases. We take proactive steps to identify and mitigate biases by using diverse and representative training data, testing for fairness, and conducting regular audits.
7. Environmental Sustainability
Recognising the environmental impact of AI, we aim to use AI efficiently and avoid unnecessary computational waste. We prioritise AI providers with strong environmental credentials and sustainable practices.
8. Stakeholder Engagement and Ethical Review
We proactively engage with stakeholders, clients, academic experts and policymakers to ensure responsible AI use. Where appropriate, we establish ethical review processes to assess AI projects for compliance with our values and legal obligations.
9. Transparency in Decision-Making
We ensure transparency in decision-making processes where AI is used:
- Clearly communicating AI involvement in decision-making.
- Maintaining records of AI-influenced decisions, including rationale and data inputs.
- Establishing accountability through human oversight and intervention where necessary.
10. Quality Assurance of AI-Generated Content
We are committed to ensuring that AI-generated content meets our quality standards:
- All AI-generated content undergoes rigorous review and quality assurance.
- Outcomes must align with our organisational values, mission and strategic goals.
- AI processes are continually assessed and refined for improvement.
11. AI Model Training & Updates
- Our models are periodically retrained on the latest data to improve accuracy and alignment with real-world conditions.
- Key performance indicators (e.g. precision, recall, and ROC-AUC) are continuously tracked to detect and mitigate model drift.
- Models undergo an approval process before deployment, ensuring they meet performance thresholds. Older versions are retained for auditability and rollback if necessary.
12. Ongoing AI Governance
- AI models are regularly evaluated and improved based on real-world performance and user feedback.
- AI usage policies are reviewed periodically to align with evolving best practices and regulatory changes.
- Customers maintain control over how AI-generated insights are applied within their workflows.
- Customers can opt out of AI features by contacting the support team.
Further Information
For any questions regarding this policy or concerns about our AI use, please contact product@convertr.io.