• The Act organizes AI systems into four categories: unacceptable risk, high risk, limited risk, and minimal risk.
• The AI Act requires that users are made aware when they're interacting with AI—especially in cases that involve content generation, biometric data processing, or classification.
• Businesses that develop or deploy high-risk AI systems must implement strict governance practices, including risk assessments, record-keeping, and registration of AI systems in an EU-managed database.

By Lisa Brown

Introduction

In a historic move, the European Union (EU) recently passed the AI Act, which serves as the first comprehensive regulation governing the development and deployment of artificial intelligence (AI). This legislation sets clear guidelines for AI use while aiming to protect fundamental rights. For stakeholders in the printing and document imaging industries, this regulatory shift brings challenges and opportunities alike as AI becomes more embedded in document workflows, content recognition, and data automation. This article provides some highlights about the EU AI Act and discusses why they matter to your business.

Categories of Risk

The Act organizes AI systems into four categories: unacceptable risk, high risk, limited risk, and minimal risk. Applications like facial recognition used in public surveillance fall under “unacceptable risk,” whereas AI in HR document processing, financial services, and legal compliance may be considered “high risk.”

This is important because many print and imaging workflows involve automated document processing, identity verification, and data extraction. These are functions that may fall under high-risk categories, depending on context. Today’s businesses must assess their solutions for compliance and determine which systems need additional oversight or certification.

Requirements for Transparency

The AI Act requires that users are made aware when they're interacting with AI—especially in cases that involve content generation, biometric data processing, or classification. This could include AI that auto-generates text, summarizes contracts, or uses optical character recognition (OCR) combined with identity verification.

Solutions that automatically generate or manipulate documents will need to clearly disclose AI involvement. If a platform helps organizations classify, redact, or route sensitive documents using AI, user transparency must be built into the user experience and any documentation.

Obligations for High-Risk Systems

Businesses that develop or deploy high-risk AI systems must implement strict governance practices, including risk assessments, record-keeping, and registration of AI systems in an EU-managed database. Fines for non-compliance are substantial; they can reach up to €35 million or 7% of global revenues.

If your print/imaging solutions are used in regulated sectors like healthcare, finance, or public services, you may be subject to these requirements. Compliance will require collaboration between technical, legal, and operational teams—and this might also become a differentiator in competitive bids.

A Real-World Example

Suppose that a financial services provider uses AI-enhanced OCR to process scanned contracts, invoices, and compliance documents. The system uses natural language processing to extract key clauses and flag potential compliance risks automatically. This kind of AI falls into a high-risk category under the AI Act due to its use in financial decision-making and legal compliance. It must meet transparency rules, meaning that users need to be aware that AI is reviewing and interpreting their documents. Without proper oversight and documentation, the organization risks significant fines for non-compliance.

Recent and Upcoming Developments

Here’s a quick breakdown of recent updates and what to expect from the EU AI Act over the next two years:

• In early February 2025, the first set of banned AI practices (like social scoring and manipulative AI) became enforceable.
• In May 2025, the EU released a Code of Practice for general-purpose AI tools, with guidance on transparency, risk, and copyright.
• On August 2, 2026, most of the main requirements for high-risk AI systems will take effect, including documentation, oversight, and registration rules.

The EU is also emphasizing energy-efficient AI and raising concerns about how AI tools use copyrighted content, signaling a push for more responsible AI use across industries.

The Bottom Line

The EU AI Act is not just a legal obligation—it’s a strategic signal. For the printing and document imaging industries, it underscores the importance of responsible AI design in automation, content recognition, and digital transformation. By proactively aligning with these standards, businesses can strengthen customer trust, reduce legal risks, and stay ahead in a rapidly evolving market. Now is the time to audit your AI capabilities, engage cross-functional teams, and begin shaping an AI roadmap that is both innovative and compliant.

As a Principal Analyst at Keypoint Intelligence, Lisa Brown has played a crucial role in enhancing client performance, leading research projects within the consulting group, and contributing to the advisory services team. She is dedicated to understanding and addressing clients' needs, developing strategies that are designed to maximize their potential. With over a decade of experience in the workplace technology sector, Ms. Brown excels in delivering actionable insights that prepare clients for success.