Guides
Ten steps to creating an AI policy
Creating your company’s AI policy involves carefully considering various ethical, legal, and operational aspects.
Here’s a ten-step guide on how a board of directors can develop an AI policy – and communicate it effectively to the executive management team and staff.
1: Establish a working group
Form a working group of board members, executives, and relevant stakeholders to lead the AI policy development process.
This group will oversee policy creation, gather necessary expertise, and ensure representation from various departments and stakeholders.
2: Educate the board
All board members should have a foundational understanding of AI and its ethical implications.
Board members should have training sessions or workshops to familiarise themselves with essential AI concepts, such as algorithmic bias, privacy concerns, and AI’s potential impact on employment.
3: Define the policy’s objectives
Identify your organisation’s primary objectives in adopting AI technology.
These objectives will shape the overall direction of the policy. This may include improving your company’s efficiency, enhancing customer experience, or promoting innovation.
4: Assess the ethical principles and values
Determine the ethical principles and values that guide AI development and deployment within your organisation.
It would help if you considered fairness, transparency, accountability, and human well-being concepts. These principles will help establish a solid ethical foundation for the AI policy.
5: Evaluate legal and regulatory compliance
Understand the legal and regulatory landscape surrounding AI, including data protection laws, privacy regulations, and industry-specific guidelines.
Ensure the AI policy meets these requirements to avoid legal risks and uphold compliance.
6: Identify potential AI use cases and risks
Identify the specific use cases and applications of AI within your organisation – where will it be used, by whom and for what purpose?
Assess the associated risks, including potential biases, security vulnerabilities, and unintended consequences. Next, develop guidelines and best practices to mitigate these risks.
7: Establish accountability and governance
Who will be responsible for your AI policy?
Define the roles and responsibilities of stakeholders involved in AI development, deployment, and monitoring.
Establish clear lines of accountability and governance mechanisms to ensure ethical decision-making and risk management throughout the AI lifecycle.
8: Ensure transparency and explainability
Promote transparency and explainability in AI systems by requiring clear documentation, responsible data practices, and understandable algorithms.
Ensure that stakeholders, including employees and customers, can comprehend the basis of AI decisions and raise concerns if necessary.
9: Encourage continuous monitoring and evaluation
Implement mechanisms to monitor an AI system’s performance, impact, and adherence to ethical standards over time.
Regularly evaluate the policy’s effectiveness and make necessary adjustments based on feedback and emerging best practices.
10: Communicate the AI policy
Craft a comprehensive AI policy document that encompasses all the elements above.
The policy should be written in clear, accessible language and provide practical guidance.
Communicate the policy approach to the executive team and staff through various channels, such as company-wide emails, town hall meetings, and training sessions.
The basic concepts of AI must inform your policy
Remember, your board’s understanding of the ethical use of AI should include critical concepts such as:
- Algorithmic bias: Awareness of the potential for AI systems to perpetuate biases and the need to mitigate such biases during development and deployment.
- Privacy and data protection: Knowledge of relevant privacy laws and regulations, understanding the risks associated with AI’s use of personal data, and ensuring compliance with data protection practices.
- Accountability: Understanding the need to establish clear lines of accountability for AI systems’ decisions and actions, including responsibility for errors or unintended consequences.
- Ethical decision-making: Familiarity with ethical frameworks and principles to guide AI development and deployment, including fairness, transparency, accountability, and human well-being.
- Social implications: Recognising the potential impact of AI on society, employment, and equity and considering measures to address these implications responsibly.
AI expertise will be helpful
A board-level understanding of AI is essential to implement its potential and mitigate its risks.
Therefore, boards need to familiarise themselves with AI technologies relevant to their sectors and their potential impact on the company’s strategic goals.
“This will help them to engage in thoughtful discussions, ask informed questions, and provide guidance on AI-related matters in the business,” says David W Duffy, CEO of the Corporate Governance Institute.
“By incorporating AI expertise at the board level, organisations can make informed decisions that align with their long-term objectives and relevant stakeholder interests.”
Your AI policy will evolve as AI advances
Remember, AI is a relatively new phenomenon that is rapidly evolving. Depending on the business, it will be of varying importance and relevance. All companies can take the following steps immediately:
- Determine how AI can be of most benefit to the business
- Assess the risks that AI could bring
- Develop an AI policy
- Appoint an AI champion, either on the board or reporting directly to the CEO
- Monitor and review your AI policy every quarter
Download an AI policy template
This AI tool usage policy template from Workable can help you draft an AI tool usage policy to ensure your organisation’s responsible and secure use of artificial intelligence (AI) tools. You can modify it based on your needs.
Download here.