In this post, we’ll highlight key points from the survey and integrate the survey to relevant items proposed by the AI Act so that you’ll be equipped with baseline knowledge to figure out your next steps. 

Here are the top highlights from the executive survey.

Rise of generative artificial intelligence

There’s no doubt most organizations are increasing their investments in genAI such as ChatGPT and recognize its potential to revolutionize productivity and customer experience. While many are exploring its benefits, full-scale implementation remains in its infancy almost across the board.

State of data and analytics

There’s a notable increase in organizations using data and analytics for business innovation, with significant improvements in managing data as a business asset. This shift reflects a growing recognition of the value data brings to the table. As we all know, your AI strategy is contingent upon your data strategy.

Challenges in data quality and talent acquisition

Despite advancements, challenges persist in data quality and talent acquisition (See this post for a different perspective on talent acquisition). Only a minority of organizations feel they have successfully improved data quality, and half report a talent gap in effectively implementing Generative AI.

Data quality and data ownership remains a challenge for many organizations. Improving data quality and management is crucial for complying with the EU AI Act’s emphasis on reliable and high-quality data for AI systems.

This is underpinned by the recent news stories about use and potential misuse of data owned by organizations such as the New York Times. 

The evolving role of the CDO/CDAO

The role of CDOs and Chief Data & Analytics Officers (CDAOs) continues to gain prominence, with more organizations acknowledging their critical responsibility for data and analytics. However, there’s still a journey towards making these roles successful and well-established across industries, and we know from other research that the average tenure of a CDO is only 18 months so something will need to change for the CDO role to have the impact that organizations desire.

Industry-specific trends

The survey indicates varied focuses of the CDO/CDAO function across different industries. For example, in less regulated sectors like retail, the focus is more on growth and innovation, whereas in regulated industries like finance, the emphasis is on compliance and risk management. Importantly, we take a look at high-risk AI systems that include facial recognition and biometric identification in medical devices and healthcare as well as insurance and financial services.

Business integration and cultural shifts

A significant number of organizations report progress in integrating data and AI into their business processes. However, the survey identifies culture, people, processes, and organization as the principal challenges in becoming data-driven, underscoring the need for cultural shifts within organizations. Knowing this, data mesh is very much still in the picture.

Survey insights guide Data and AI Leaders navigate the proposed AI regulation

The survey’s insights are critical to CDOs and Ai leaders, especially with the upcoming EU AI Act. Understanding these trends and challenges help create effective strategies that comply with the upcoming regulation that harness the potential of AI.

6 things executives need to know as it relates to the upcoming EU AI Act

Based on the survey results above, here are six things executives need to know with an eye on the upcoming EU AI Act. These insights from the survey provide a roadmap for CDOs to align their strategies with the proposed EU law, focusing on ethical, responsible, and effective use of AI systems.

1. Safeguards and guardrails for AI

The EU AI Act emphasizes the need for ethical AI practices for unacceptable risk. The survey reveals that 99% of respondents believe in the necessity of safeguards for Generative AI, aligning with the Act’s focus on AI governance.

2. High-risk AI systems checklist | Which AI systems will the Act cover?

High-risk AI systems include: medical devices, recruitment and HR-based apps, financial and insurance services and providers, facial recognition or biometric identification, border control, social scoring, and vehicles. 

Key requirements vital to creating a safe and secure general purpose AI system include:

• Creating a fundamental rights impact assessment
• Registering in the public EU database for high-risk AI systems
• Implementing risk management system (e.g. It categorizes AI systems based on risk levels, focusing particularly on high-risk AI applications)
• Transparency obligations (e.g. The Act emphasizes transparency in AI systems, requiring clear information to users)
• Data governance (e.g. addresses the quality and integrity of data used in AI systems)
• Human oversight (e.g. explainability, audit logs, human-in-the-loop)
• Accuracy and cybersecurity (e.g. continuous monitoring and testing)

3. Data ethics and corporate responsibility

The survey indicates that a significant number of organizations prioritize data ethics. Aligning with the EU AI Act’s ethical considerations is essential. Ethical considerations in AI usage have gained momentum especially with problematic deepfakes. 

Most organizations acknowledge the need for robust policies and practices. The survey highlights a collective call to action for the tech industry to address data and AI ethics more vigorously.

4. Understanding AI’s business impact while maintaining compliance

Understanding and articulating the business value of AI investments is crucial, especially as the EU AI Act may require justifications for AI use in certain applications. Integrating AI into business processes is a key trend. The EU AI Act will require processes to be compliant, transparent, and responsible.

5. Risk management and AI governance

The survey notes concerns about risks like misinformation and job displacement due to AI. The EU AI Act policymakers and lawmakers focus on the level of risk via a risk assessment and AI governance aligns with these concerns, emphasizing the need for effective risk management strategies in AI deployment.

6. What are the penalties for non-compliance?

Non-compliance with this regulation will of course mean fines and they’re substantial:

Administrative fines

Fines can be up to €30,000,000 or up to 6% of the company’s total worldwide annual turnover for the preceding financial year, whichever is higher.

Other forms of non-compliance

Fines can reach up to €20,000,000 or up to 4% of the company’s total worldwide annual turnover for the preceding financial year, whichever is higher.

False or misleading information to regulatory bodies or national competent authorities

The fines can be up to €10,000,000 or up to 2% of the company’s total worldwide annual turnover for the preceding financial year, whichever is higher​​.

Conclusion: With great promise comes great responsibility

The 2024 Data and AI Leadership Executive Survey paints a vivid picture of a rapidly evolving landscape. Generative AI emerges as a game-changer, yet its full potential is still to be harnessed. CDOs and CDAOs are increasingly pivotal, yet their roles continue to evolve amidst challenges in data quality, talent, and ethical AI governance. 

As we look ahead, leaders must navigate these complexities, ensuring their strategies are aligned with the latest trends and regulatory landscapes, including the proposed Artificial Intelligence Act. The journey towards a data-driven and trustworthy AI will evolve, with much promise and work ahead to mitigate systemic risks for lawmakers, business executives, and dare we say: even data engineers.