Bridging the gap: Improving public service communication through AI

Most citizens have likely stared at an official form or government website and felt as if it were written in a foreign language. Archaic words, complex sentences, and legal jargon make public documents inaccessible to many, creating a silent barrier between citizens and the services they need. Yet, over time, people tend to accept it as part of dealing with public institutions—when it shouldn’t be.  

For decades, institutions have tried to address the problem. However, progress has been slow, especially as manual simplification of texts is laborious and difficult to scale, leaving millions of users to navigate content that is confusing, dense, and often intimidating. 

Why bureaucratic language is still so hard to understand 

Despite decades of reform efforts, bureaucratic language remains notoriously difficult to understand. For citizens of all ages and backgrounds, navigating official documents or government websites often proves unnecessarily complex and time-consuming. 

The reasons behind this are many.  

Some include:  

1. Legalistic and formal language: Many official texts prioritize legal precision over clarity. Words and phrases are often chosen for their technical or legal correctness, even if they confuse the average reader. This is especially true in regulations, contracts, and formal announcements. 
2. Archaic terms and traditional conventions: Administrative language tends to rely on terminology that has been used for decades — sometimes centuries — without modernization. Phrases like “hereinafter referred to as” or “pursuant to the provisions of” are common, making the text feel distant and intimidating. 
3. Overly long sentences and dense structure: Bureaucratic writing often chains multiple clauses together, creating sentences that span several lines. This density increases cognitive load, making it harder to extract key information quickly. 
4. Fragmented information across multiple documents: Instructions or procedures are rarely consolidated in a single place. Citizens may need to cross-reference multiple forms, laws, or online portals, further complicating comprehension. 
5. Inconsistent formatting and presentation: Even when information is technically clear, poor layout, inconsistent headings, or lack of summaries can make it inaccessible. Dense, unstructured text slows readers down and leads to more misunderstandings.  

These challenges aren’t abstract — they have tangible effects on everyday life. Citizens may misinterpret procedural instructions and miss important deadlines, find it difficult to access essential benefits or services, and feel excluded from participating in civic life or decision-making. Over time, such experiences can erode trust in institutions that come across as opaque or indifferent. 

Why simplification has been slow 

Efforts to modernize bureaucratic language often face systemic hurdles: 

• Limited resources: Public administrations may lack staff trained in both subject matter and plain-language communication. 

• Manual workload: Simplifying complex documents by hand is laborious, making it difficult to scale across hundreds of services. 

• Fear of legal misinterpretation: Officials may resist simplification for fear of inadvertently changing the legal meaning of a document. 

However, AI is changing all of this. Advanced language technologies are now able to tackle the complexity of bureaucratic texts at scale, offering solutions that are not only faster but also more consistent and precise than traditional manual methods. 

AI is increasingly viewed as a tool to help make bureaucratic language clearer and more accessible to people. 

More specifically, Composite AI combines different technologies — such as natural language processing, semantic analysis, and machine learning — to ensure results are accurate, consistent, and legally sound. Generative AI, on the other hand, can rephrase and summarize complex documents. 

Key ways AI supports accessibility include:  

• Automated text simplification: Generative AI can rewrite complex documents into clearer, more accessible language. The system preserves the original meaning while improving readability, helping citizens navigate services more easily. 

• Objective readability measurement: AI can apply quantitative metrics such as the Gulpease Index to evaluate sentence and word complexity. Scores range from 0 to 100, with higher values indicating easier reading. Using such metrics allows administrations to track accessibility improvements and verify that simplified texts are genuinely easier to understand. 

• Scalability across services: AI enables public administrations to simplify hundreds of documents efficiently, moving beyond slow, resource-intensive manual editing and supporting a consistent, accessible communication strategy across multiple services. 

• Human oversight: While AI generates drafts, human experts review all edits to ensure accuracy, legal compliance, and contextual appropriateness. This workflow combines automation with professional oversight to maintain high-quality outcomes. 

Simplifying administrative language presents both practical and conceptual challenges. Public institutions must ensure that their communication is accessible and inclusive, without compromising legal precision or institutional integrity.

Giacomo Grassi, Director of User Experience at INPS launched a large-scale initiative, in collaboration with Almawave, to promote clear and accessible communication for all citizens. Its digital ecosystem serves millions of users and delivers hundreds of services through thousands of online applications.

The complexity of bureaucratic language often acts as a barrier, limiting people’s ability to understand procedures, access services, and fully exercise their rights. Simplification, however, goes beyond rewriting: it requires a structured, legally sound methodology that balances linguistic clarity with procedural accuracy.

To achieve this, the organization introduced dedicated writing guidelines, review workflows, and measurable readability indicators to ensure that each simplified text remains transparent, consistent, and legally valid. Early results show a significant increase in accessibility, with millions of additional users now able to understand public information more easily.

More recently, the institution has also begun experimenting with generative AI to support the simplification process. The AI produces first drafts that are reviewed and refined by human experts, then compared with manually simplified versions. Large-scale tests confirm that simplified texts—whether human-edited or AI-assisted—are consistently preferred for their clarity, reliability, and ease of reading.

This experience highlights a broader principle: improving linguistic accessibility is not just a matter of communication strategy, but one of equity and inclusion. When public information becomes easier to understand, institutions strengthen trust and foster more open, democratic participation.

Accessibility as a right 

Making bureaucratic language accessible is about more than efficiency or convenience — it is a matter of digital inclusion and citizen empowerment. Simplifying language ensures that everyone, regardless of education, background, or ability, can understand and exercise their rights—a step toward ethical and inclusive governance. 

Beyond technological innovation, Almawave’s work with public institutions is grounded in compliance by design—embedding regulatory adherence and ethical principles into every stage of AI development and deployment. 

Almawave holds ISO/IEC 42001 certification, the international standard for AI management systems, demonstrating systematic and responsible AI development with clear accountability and risk management processes. 

Additionally, Almawave is a signatory to the Code of Conduct for General Purpose AI, affirming its commitment to transparency, fairness, and responsible innovation in line with European AI regulations. 

This ensures that public administrations can trust their digital transformation initiatives are not only effective but also ethically sound, legally compliant, and aligned with citizen protection principles—essential for maintaining public trust as AI reshapes public services. 

Discover how AI and human expertise can make institutions more open and inclusive.