Designing Secure and Trustworthy AI Platforms: Privacy and Responsibility in Early Stage Product Development

Summary

Our event, Designing Secured and Trustworthy AI Platforms: Privacy and Responsibility in Early Stage Product Development, was created to help early career developers, product managers, start-up founders, and technology professionals understand how to build AI products that are secure, trustworthy, and responsible from the very beginning.

Many early stage teams are under pressure to build quickly, attract investment, and release products as fast as possible. As a result, important issues such as security, privacy, transparency, and responsible innovation are often treated as tasks to complete after a product has been launched. Through this training, we wanted to encourage participants to adopt a different mindset by embedding these principles into product design and development from day one.

The event brought together experts from industry and academia to deliver practical and interactive sessions. The programme began with an introduction to JournoTECH and NewsAssist AI, demonstrating how responsible AI can be applied in practice across journalism, research, and other sectors. Participants then explored human centred product design, learning how to make intentional design choices by clearly identifying who a product is designed for while considering everyone who may be affected by it, including end users, regulators, executives, and other stakeholders.

The second day focused on secure cloud architecture. Participants gained practical knowledge of how cloud systems are structured and how to protect them using principles such as Zero Trust, Identity and Access Management, secure data storage, access controls, firewall rules, and backup and recovery planning. Complex technical concepts were explained in an accessible way, enabling participants with different levels of experience to understand how security can be built into AI platforms from the outset.

The event created opportunities for participants to reflect on how design decisions affect privacy, security, trust, and society. Discussions encouraged participants to think beyond technical functionality and consider the broader consequences of AI systems, particularly for journalists, activists, and communities that are often disproportionately affected by poorly designed technologies.

Participant feedback demonstrated that the training achieved its objectives. Many reported that they had shifted from a technology first mindset to one that places people, security, and responsible innovation at the centre of product development. Several participants said they would redefine their target users, avoid making design decisions based on assumptions, and integrate security into their products from the earliest stages rather than treating it as a final step.

The event also generated valuable insights into the challenges faced by early stage developers when balancing innovation with privacy, trust, and security. These discussions will inform JournoTECH's future work in developing practical guidance and training resources for responsible AI development.

Participants also expressed strong interest in future training on secure AI development, Large Language Models, secure API integration, and advanced cloud security, demonstrating continued demand for this work.

Highlights

Building AI that people can trust starts before the first line of code

What if security and privacy were not problems to fi x after launch, but principles that shaped every design decision from the beginning?

That question was at the heart of JournoTECH's two day training, Designing Secured and Trustworthy AI Platforms: Privacy and Responsibility in Early Stage Product Development, supported by SPRITE+.

The event brought together early career developers, product managers, start-up teams, and technology professionals to explore how trustworthy AI can be built from day one. Through practical sessions led by experts from academia and industry, participants explored human centred product design, responsible innovation, and secure cloud architecture.

The training challenged participants to think differently about product development. Rather than trying to build products for everyone, they learned the importance of identifying the right users, understanding who may be affected by design decisions, and considering security and privacy throughout the development process.

The feedback showed that the learning translated into action. Participants said they would rethink their approach to product design, defi ne their users more carefully, avoid making assumptions, and prioritise security from the earliest stages of development. Others highlighted how the training helped them better understand cloud security, Identity and Access Management, access controls, and protecting data throughout the product lifecycle.

The event demonstrated that responsible innovation is not only about building smarter technology. It is about building technology that people can trust. By equipping the next generation of developers with practical knowledge and responsible design principles, JournoTECH is helping shape a future where security, privacy, and accountability are built into AI from the very beginning.

Impact

The project generated meaningful social, organisational, and sector wide impact by strengthening participants' understanding of how responsible AI can be embedded into product development from the earliest stages.

Socially, the event encouraged participants to design technology that is more inclusive, accessible, and accountable. Rather than focusing only on end users, participants learned to consider everyone affected by a product, including regulators, decision makers, and communities that may experience unintended harm. This shift in thinking supports the development of AI systems that are more trustworthy and equitable.

From an organisational perspective, participants reported that they would change how they develop products by defining their target users more clearly, avoiding assumptions during design, and integrating security and privacy from the outset. They also gained practical knowledge of cloud security, access management, data protection, and system resilience, helping them make better technical and strategic decisions within their organisations.

The project also has economic value. Building secure and privacy conscious systems from the beginning can reduce the costs associated with redesign, security breaches, and regulatory compliance later in the product lifecycle. This is particularly valuable for start-ups and early stage teams operating with limited resources.

At a wider sector level, the discussions generated qualitative insights into the challenges developers face when balancing rapid innovation with responsible practice. These findings will inform future JournoTECH training, contribute to responsible AI conversations, and support the wider goal of promoting security, transparency, and accountability across the AI ecosystem. The strong demand for additional training demonstrates that there is a clear appetite within the developer community for practical guidance on trustworthy AI development.

Lessons Learned

The event reinforced the importance of combining technical expertise with practical discussion and reflection. Participants responded positively when complex concepts were explained using real world examples and when they were encouraged to relate the learning directly to their own products and professional experiences.

One important lesson was that developers are eager to learn about responsible AI but often have limited opportunities to explore these topics in practical ways. The discussions showed that many participants had previously viewed security and privacy as technical or compliance issues rather than core design principles. Creating space for conversation alongside technical instruction helped participants develop a more holistic understanding of trustworthy AI.

The feedback also highlighted the value of bringing together speakers from different disciplines. Combining expertise in product design, cloud architecture, and responsible innovation enabled participants to see how technical, ethical, and organisational considerations are interconnected.

Another lesson was the demand for more hands on learning. Participants requested future sessions on secure AI platforms, Large Language Models, API integration, and practical implementation of AI systems. This suggests future programmes would benefit from additional workshops, demonstrations, and applied exercises that allow participants to put concepts into practice.

Outcomes & outputs

The event successfully delivered a range of tangible outputs and outcomes that align with the objectives of the project.

The programme provided early career developers, product managers, start-up founders, and technology professionals with practical training on responsible AI, human centred product design, and secure cloud architecture. Participants gained a stronger understanding of how to integrate security, privacy, transparency, and accountability into AI platforms from the earliest stages of development.

The event facilitated interdisciplinary knowledge exchange by bringing together experts from industry and academia to share practical experiences and encourage discussion around responsible innovation. Participants explored the relationship between product design, cloud infrastructure, user experience, privacy, and security, helping them understand how these areas influence one another.

Participant feedback demonstrated measurable learning outcomes. Many reported changing how they think about product development by placing greater emphasis on understanding users, designing for broader stakeholder groups, and avoiding assumptions during the design process. Participants also indicated that they would integrate security considerations from the beginning of development rather than treating them as activities to be completed after deployment.

The secure cloud architecture session increased participants' understanding of topics including Zero Trust principles, Identity and Access Management, secure data storage, access control, firewall rules, cloud infrastructure, backup strategies, and system resilience. Several participants noted that the session made complex technical concepts easier to understand and apply.

The event also generated qualitative evidence on the challenges faced by early stage developers when balancing rapid innovation with security, privacy, and responsible AI practices. These insights will help inform future JournoTECH programmes and contribute to wider discussions on trustworthy AI development.

Evaluation findings demonstrated high participant satisfaction and continued engagement. All respondents requested that the postponed final session be rescheduled, indicating strong interest in completing the programme. Participants also identified priority areas for future training, including designing trustworthy AI platforms, practical implementation of Large Language Models, secure API integration, advanced cloud security, and deeper exploration of NewsAssist AI.

As a result of the project, JournoTECH has strengthened relationships with developers, start-up teams, academic partners, and industry experts while expanding its growing community of practitioners interested in responsible AI. The event has also created a foundation for future workshops, practical guidance, and collaborative activities that support secure, trustworthy, and responsible AI innovation.

Overall, the project successfully met its objectives by increasing awareness, improving practical knowledge, encouraging behavioural change, generating valuable qualitative insights, and creating opportunities for continued engagement around responsible AI development.