Responsible AI roadmap
“AI can be a force for good that helps to solve critical challenges. However, it may also impact our fundamental human rights. This creates a responsibility for companies that create and use these technologies to act ethically. The mobile industry is committed to connecting everyone to a better future, and is well placed to contribute to this evolving field.” Mats Granryd, Director General of the GSMA.
Exercise: Take a few minutes to consider where you are in your responsible AI roadmap. Use this checklist not just as a step-by-step guide, but as a tool to identify gaps in your practices and determine key focus areas for future development. This will help you strategise effectively on what actions to take next in advancing your responsible AI initiatives.
Principles: Establishing foundations for responsible AI by design
- Assess and identify existing processes that could be integral to the implementation of responsible AI, such as leveraging data privacy principles, data governance frameworks from GDPR, and security protocols, among others.
- Formulate and set forth clear AI ethical principles and guidelines to guide the implementation of responsible AI by design.
Governance: Establishing teams and leadership for responsible AI oversight
- Initiate an educational programme focused on responsible AI for executive-level management, promoting a top-down approach to ethical AI implementation.
- Establish an executive committee or advisory panel to drive the initial understanding, advise, and propose practices and policies for responsible AI by design. Ensure that this panel includes representation from various departments to foster diverse perspectives.
- Implement a risk management system focused on minimising bottlenecks in AI deployment, while ensuring that each solution receives the appropriate level of support and oversight based on its specific risk profile. This approach streamlines processes and effectively addresses unique AI-related risks for comprehensive risk mitigation.
- Create new responsibilities (such as champions or experts) or form a dedicated team to assist AI product managers in the implementation of responsible AI, providing necessary support and guidance.
Processes: Enhancing AI ethics through training, communication, and workforce diversity
- Implement an awareness programme to ensure organisation-wide understanding of the importance of responsible AI by design. This program should educate all levels of staff about the principles and responsible AI practices.
- Implement a strategy to regularly communicate the value of responsible AI, ensuring clear understanding across all organisational levels.
- Equip product managers with training in responsible AI, empowering them to integrate responsible AI practices into every stage of product design, development, and deployment, ensuring ethical considerations are embedded from the outset.
- Implement a tool, such as the self-assessment questionnaire (SAQ) to establish and address potential risks when initiating a new AI project.
- Establish a change management process to smoothly integrate the adoption of ethical principles, especially if this represents a significant shift from the current business operations.
- Actively study concrete examples and monitor ongoing ethics risks associated with AI systems as their scale and scope increase to continually adapt and improve ethical practices.
- Encourage workforce diversity in AI teams to bring varied perspectives.
Thought leadership: Fostering engagement in industry and regulatory activism, and promoting responsible AI in external communications
- Actively connect with external experts to exchange best practices and gain insights into the evolving field of responsible AI.
- Engage in industry and regulatory activism, and actively participate in discussions about AI ethics in external communications to contribute to the broader dialogue and shape the future of ethical AI practices.
Independent Oversight: Ensuring compliance through third-party responsible AI audits or assessments
- Actively engage in AI ethics audits or assessments to ensure compliance with ethical standards and to gain an objective evaluation of your AI practices and policies.
“stc is focused on the introduction of innovative AI products and solutions to produce correct, precise and reliable results insights for our customers. We are delighted to collaborate with the GSMA and the other leading operators in developing “International Ethical AI Principles” for use in the telecom industry, and correspondingly embedding into stc’s culture. This demonstrates stc’s practice of complying to Respect of Privacy, Transparency and Openness, and contributing to international standards.
By adhering to that, we believe it will not only enhance our customers experience and trust, but will also contribute to resource optimisation, improving the efficiency of our products development and deployment, and supporting data/model re-usability. This is an exciting milestone in stc’s digital transformation journey, stc will continue to endeavour to always lead the market towards digital transformation and the localisation of AI.”
Eng. Haithem AlFaraj, Chief Technology Officer, Saudi Telecommunications Company (stc)