Introducing AI applications

This high-level guideline is intended for senior parliamentary managers, as well as for parliamentary staff and MPs who are interested in gaining a broad understanding of where AI can impact upon the work of parliaments.

This guideline outlines considerations and recommendations for parliaments facing new challenges in procuring, implementing and managing AI systems at a time when this technology is becoming increasingly prevalent in standard software packages. It addresses how parliaments can approach AI-enhanced off-the-shelf products while adhering to the ethical principles and risk management strategies outlined in these Guidelines.

Increasingly, many off-the-shelf software packages are being augmented with AI functionalities. This AI integration is often opaque, and it is not immediately apparent to the user what impact AI is having, or how it works behind the scenes:

• Microsoft 365 and Microsoft Edge products are starting to embed Microsoft Co-pilot AI support.
• Google Docs already has new AI features such as “Help me write”, “Smart compose”, “Summarization” and “Voice typing”.
• Adobe Acrobat has integrated several AI functionalities, including AI Assistant for generating summaries and creating multi-document insights.
• Photo and video editing software often contains AI augmentation that makes it easy to render manipulated images.

When procuring AI-enhanced products, parliaments must therefore exercise due diligence to ensure alignment with their ethical standards and operational needs.

Checklist:

• Conduct thorough vendor assessments, focusing on AI ethics and data practices.
• Evaluate AI features against parliamentary needs and ethical guidelines.
• Ensure contracts include robust clauses on data protection and AI accountability.

A measured approach to implementation allows parliaments to assess the impact of AI functionality and its alignment with existing processes. Starting with a pilot phase provides an opportunity to develop clear protocols for AI feature usage and establish monitoring mechanisms.

Checklist:

• Begin with a pilot phase to evaluate AI impact and alignment.
• Develop protocols for enabling or restricting AI features based on task sensitivity.
• Establish monitoring mechanisms for AI-driven decisions or suggestions.

Comprehensive training is crucial to ensure staff can leverage AI features effectively while understanding their limitations and potential risks. This is particularly important given the ethical considerations and potential biases inherent in AI systems.

Checklist:

• Develop training programmes highlighting AI benefits and risks (including data literacy and AI literacy).
• Create user guidelines for the appropriate usage of AI functionalities.
• Educate staff on recognizing and critically evaluating AI-generated content.

As AI capabilities evolve, parliaments must regularly review and update their policies and practices. This includes maintaining open communication with vendors, conducting periodic audits, and ensuring ongoing alignment with ethical principles such as fairness, transparency and human oversight.

Checklist:

• Regularly review and update AI usage policies.
• Maintain open communication channels with vendors.
• Conduct periodic audits of AI feature usage and impact.
• Assess and mitigate potential biases in AI-enhanced features.
• Ensure transparency and maintain human oversight in AI use.

Strict data governance is essential when using AI-enhanced products, particularly given the sensitive nature of parliamentary work. Regular performance evaluations help to ensure that AI features continue to meet parliamentary needs and ethical standards.

Checklist:

• Implement strict data access controls for AI-enhanced features.
• Ensure compliance with data protection regulations and parliamentary policies.
• Regularly assess the effectiveness and appropriateness of AI features.
• Gather user feedback to inform future procurement and implementation decisions.

By following these recommendations and checklists, parliaments can harness the benefits of AI-enhanced off-the-shelf products while mitigating risks and upholding ethical standards. This approach aligns with the broader AI governance framework outlined in the Guidelines, ensuring a consistent and responsible approach to AI adoption across all areas of parliamentary work.

The Guidelines for AI in parliaments are published by the IPU in collaboration with the Parliamentary Data Science Hub in the IPU’s Centre for Innovation in Parliament. This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International licence. It may be freely shared and reused with acknowledgement of the IPU. For more information about the IPU’s work on artificial intelligence, please visit www.ipu.org/AI or contact innovationipu [dot] org.

About the guidelines | The role of AI in parliaments | Inter-parliamentary cooperation for AI | Strategic actions towards AI governance | Risks and challenges for parliaments | Generic risks and biases | Ethical principles | Risk management | Alignment with national and international AI frameworks and standards | Project portfolio management | Data governance | Systems development | Security management | Training for Data Literacy and AI Literacy | Glossary of terms