Shaping the Future of AI
 

The EU AI Act is a new comprehensive regulatory framework that aims to ensure that artificial intelligence (AI) systems are safe and respectful of privacy laws and individuals' fundamental rights. Set to become law between May – July 2024, the Act specifies precise requirements and obligations for AI developers and deployers that aim to reduce administrative and business burdens, especially for small and medium-sized enterprises.
 

AI systems are categorized into four levels of risk, depicted in Figure 1. High-risk systems, including those used in critical infrastructure or law enforcement, must comply with strict requirements before deployment. Low or minimal risk systems like chatbots will have lighter regulatory burdens. 

The new rules aim to foster trustworthy AI not just in Europe but also globally. This means non-EU countries and companies wishing to operate in the European Union market must also examine and potentially adjust their AI strategies. (You can read more about the EU AI Act in our article,” EU AI Act: How to Create an Effective Data Governance Strategy for Your Organization).”
 

Navigating Data Governance Through the EU AI Act

As businesses strive to align with the EU AI Act, they are encountering various challenges, both from within and outside their organizations. Internally, they must navigate multiple teams, business units and departments to effectively manage data silos and establish a robust foundation for AI algorithms. This requires addressing risk factors confidently and consistently, adding layers of unintended complexity.

At the same time, external challenges are arising for organizations due to the uncertainty surrounding the EU AI Act and its measures. It is essential to navigate the procedures, approaches and strategies for regulatory compliance, ensuring that they have the correct data and foundation for agile, scalable data management. 

Having high-quality, reliable data is a cornerstone of effective decision-making in any modern organization; thus, robust AI governance becomes essential, focused on some key considerations (read more here): 

1. AI/ML Data Cataloging: Capture and document current AI/ML formulas and related data; assess the risk and outline a seamless process for cataloging new AI/ML methodologies.
2. Critical Datasets: Identify trustworthy and reusable datasets, curating them for AI/ML application, extending to recurrently used AI/ML data and data products over time(discover how data products can transform your company's approach to data, bridging the gap between data producers and consumers).
3. Data Quality & Observability: Evaluate data quality for data-feeding AI/ML algorithms and track critical metrics systematically.
4. Data Classification: Record data sets and traits; identify crucial processes to evaluate potential impact based on the EU AI Act.
5. Metadata & Data Lineage: Facilitate organization-wide understanding of the metadata and data lineage to evaluate any AI/ML loose ends.
6. AI/ML Data Governance: Utilize AI/ML for EU AI Act compliance with proper data and algorithm management, guided by a cross-domain team reviewing data and AI/ML strategies.
    

Integrated Cloud Data Management for the AI Era

Navigating AI data governance requires a modern approach supported by the right platform. Among data leaders implementing or planning to implement generative AI, the quality of data (42%) is the top data-related obstacle to the adoption of generative AI and large language models (LLMs).  And data sources are expanding as well: 41% admit they’re juggling 1,000 or more data sources — a number 79% expect to increase in 2024,¹  reveals an Informatica survey of 600 data leaders from large enterprises around the world.

This reflects some of the crucial pillars for modern data governance including managing risk and compliance, data sharing and democratization, and intelligent data observability. The evolving nature of generative AI raises regulatory and legal challenges. Businesses must stay abreast of relevant laws and regulations, such as intellectual property rights and data protection, to ensure compliance and mitigate legal risks. Robust data governance that ensures the availability and security of data throughout the enterprise is essential. Informatica Intelligent Data Management Cloud^™ (IDMC) provides an AI-powered, automation-driven ecosystem to support your AI/ML requirements, ensuring transparency, reliability and integrity of your data. Leveraging Informatica’s metadata-driven AI engine, CLAIRE^® (cloud-scale AI-powered real-time engine), organizations can automate various aspects of data management — such as data cataloging and lineage, data quality, data observability, master data management (MDM), data security and privacy, and data sharing —  that play a crucial role in enabling organizations to support ethical, fair, transparent and responsible AI systems. CLAIRE enhances compliance with regulations like the EU AI Act and provides businesses with a unified, AI-powered solution that delivers reliable data for intelligent decision-making.

A U.S. biopharmaceutical company with significant operations across the EU provides a real example of data governance in the AI era using IDMC. Transitioning from a rigid data management setup to a more flexible data mesh architecture, the company leveraged IDMC for this shift. This change not only enhanced the company's data management capabilities to support its strategic goal of delivering 10+ transformative therapies to patients by 2030, but also ensured compliance with the EU AI Act. 

Moreover, TelevisaUnivision modernized its on-premises data estate to a cloud-native environment using IDMC, and enhanced its data services for every employee building descriptive, prescriptive and predictive analytics models. These advancements help support their real-time decision-making and accelerate delivery of services to market. This resulted in 44% integration cost-saving and 35% increased customer service efficiency.

The business opportunity for the use of generative AI is compelling, drawing more and more companies to seek out ways to take advantage of this new technology. By 2027, spending on AI software will grow to $297.9 billion with a CAGR of 19.1%. Over the next five years, the market growth will accelerate from 17.8% to reach 20.4% in 2027. Generative AI software spend will rise from 8% of AI software in 2023 to 35% by 2027. As such, Informatica's CDO Insights 2024 Survey reveals that almost half of chief data officers (CDOs) would consider upskilling or reskilling their staff on AI and machine learning, illustrating the importance of building trust and efficacy through better AI strategies within a business and its customer interactions.
 

Looking Ahead

The EU AI Act is leading the way in AI regulatory controls, and other countries are following suit with similar initiatives such as the Singapore Model AI Governance Framework and the USA Executive Order on AI. It's clear that the future of AI globally will be guided by these principles of transparency, accountability and ethics. This reaffirms the importance of understanding and engaging with international regulations in an increasingly complex world. 

Organizations are under pressure to plan for mitigating and developing a strong solution to address the risks associated with generative AI. According to a recent survey by McKinsey, only slightly over 20% of companies have risk policies in place for this technology.²

As businesses across the world prepare for this new regulatory landscape, Informatica stands ready to assist with data and AI governance solutions that put trust, privacy and compliance at the forefront. 

Informatica remains optimistic that businesses will be able to successfully navigate these challenges and emerge stronger and better prepared for the future. By working together to establish best practices, a more stable and secure AI ecosystem can be created for all.

To learn more about how Informatica helps organizations with their AI strategy, visit us at www.informatica.com/data-governance.

1 CDO Insights 2024: Charting a Course to AI Readiness

2 Quantum Black AI by McKinsey, The state of AI in 2023: Generative AI’s breakout year