Event Recap - Data & AI Meetup: Ethics

This month, the theme in The Data Lab Community is Ethics. With data & AI technologies becoming ever more ingrained in how we live, work and play, it is crucial that fairness, transparency and accountability are at the forefront of decision making.  

The responsible development of data & AI technologies not only mitigates risks and harms to society but also fosters trust among users, promoting social good and aligning technology with human values.  

To kick things off, we hosted our monthly meetup at Bayes Centre on the 4th March and were delighted to be joined by Leonardo Bezerra, lecturer in AI and Data Science at the University of Stirling, and a member of the Participatory Harm Auditing Workbenches and Methodologies project (PHAWM), and Wiktoria Kulik, Responsible AI Manager at Accenture.  

Below you'll find a summary of the two presentations from Leonardo and Wiktoria on their respective work in data and AI ethics – make sure to check out the recording for the full event including the panel discussion and audience Q&A! 

Leonardo Bezerra – Supporting Society to Assess AI 

Artificial Intelligence (AI) is rapidly evolving, and with it, new regulations are emerging to ensure AI systems align with societal values and human rights. Leonardo’s talk focused on the importance of participatory harm auditing, particularly in light of the recently approved EU AI Act. But what is participatory harm auditing?

Participatory Harm Auditing (PHA) is a process designed to empower diverse stakeholders, including non-technical individuals, to assess AI systems for potential harm. Traditionally, AI risk assessments occur late in the development cycle, but PHA aims to introduce a more proactive approach—similar to privacy by design in GDPR—by integrating harm assessment earlier in the AI lifecycle, and crucially, involving non-AI experts.  

The PHAWM project (Participatory Harm Auditing Workbenches and Methodologies) is a collaboration among seven universities across the UK. Its mission is to create tools and methodologies that allow a wide range of participants to audit predictive and generative AI systems both individually and collectively. This approach ensures AI accountability through diverse perspectives rather than relying solely on technical experts. 

Throughout his talk, Leonardo discussed the challenges in AI assessment, bridging the gap between AI experts and society and the importance of ensuring that assessments actively involve diverse voices, and reflect societal concerns. 

Wiktoria Kulik – The Challenges of Digital Ethics and Responsible AI 

Wiktoria, a Responsible AI Manager at Accenture, discussed the evolving landscape of AI ethics and governance. With a background in digital ethics and policy, she highlighted the shift in attitudes from merely acknowledging AI’s ethical concerns to actively addressing them. The rapid rise of AI in the past few years (particularly generative AI) has understandably led to both exaggerated fears amongst the public, and real-world challenges that do, or could, cause harm to many in society. This is why responsible AI implementation must be a priority for organisations. 

Wiktoria framed the current state of responsible AI into 6 key areas: 

Ownership and accountability in AI ethics have evolved, moving from isolated efforts by well-meaning individuals, to an integrated approach involving compliance, risk, and technical teams across  organisations. Regulations have played a crucial role in accelerating responsible AI adoption by unlocking budgets and enforcing accountability. However, effective implementation requires balancing compliance with innovation, ensuring that ethical AI practices are embedded into development without hindering progress. Governance structures must be proactive, integrating ethical considerations from the early stages of AI development rather than addressing issues too late in the development cycle.  

Beyond regulations and governance, organisational culture is a critical factor in responsible AI. Wiktoria highlighted how internal structures, such as performance evaluation metrics, can discourage transparency and hinder ethical AI practices. A successful responsible AI strategy requires fostering a culture of openness, collaboration, and continuous assessment. Ultimately, organisations must balance speed and responsibility, ensuring that AI-driven advancements align with ethical principles while maintaining competitiveness in an evolving landscape.

More events coming up this month 🎟️

As part of our activities throughout the month, we're hosting and supporting a number of events related to our theme. Make sure to register for the events you'd like to attend via the links below! 

• Ethical Data Security: Are You Protecting What Matters? - 18 March 2025 12:00 - 13:00 GMT [Online] SUPPORTED EVENT

Join the Cyber and Fraud Centre to explore organisations' ethical responsibilities in handling data securely. We'll cover where businesses go wrong, what happens when data is compromised, and how to build a strong security framework. Attendees will leave with practical steps to strengthen their defences against cyber threats and ensure they manage data responsibly and ethically. 

• Responsible AI and the Role of Participatory AI-Auditing - 21 March 2025 12:30 - 13:30 GMT [Online]

AI auditing is the practice of testing whether an AI system is working as it was intended. AI auditing involves reviewing how AI makes decisions, checking whether decisions are fair, accurate and safe, and ensuring they follow ethical guidelines. In the same way businesses get audited for their finances, AI systems are audited to ensure they are not biased, unreliable or acting unexpectedly. Join the Participatory Harm Auditing Workbenches and Methodologies (PHAWM) team, who will share their thinking and reflection from their ongoing work to create co-designed tools and methodologies to audit the harms and fairness of predictive and generative AI across a range of sectors. 

• AI Ethics and Compliance: What's Changing and Why it Matters - 25 March 2025 12:30 - 13:30 GMT [Online]

Discussions on regulation and ethics in data & AI have been a growing debate as AI and the public's understanding of these technologies has quickly evolved. In this session, Michaela D'Mello, Data Specialist at LEGO Group, will link current regulation discussions to data ethics concerns and highlight the relevance of data ethics and regulation to all of us, not just those who work in ethics or compliance. Whether you have 'ethics' in your job title or not, you can still play your part in ensuring data and AI technologies are built and implemented responsibly.  

Head to the events board for the full calendar of upcoming events! And you can watch recordings of previous Data Lab Community events on The Data Lab's YouTube Channel.