Scaling Responsible Artificial Intelligence Solutions

The SRAIS project gave us the unique opportunity to deepen and advance the state-of-the-art in the standardisation methods for deploying Responsible AI-powered smart clothes, namely wearables, applied to future workers in industry and healthcare application domains. We have been able to define the steps towards a framework that manages all risks associated with the concept, design, and deployment of AI-powered wearables that predict health indicators of future workers, without impacting their privacy and while guaranteeing reliable outcomes. Now the problem is to understand how the framework may lead to clear regulations impacting Italy and Europe, while ensuring a level of compatibility with the international standards.

We learned many lessons from these mentorship experiences that we hope can be useful for other teams and mentors as well. First, mentors need to be aware that teams face diverse challenges while scaling responsibly. It's important to be open to find the most pressing challenge and focus on that. Second, even a few months of mentorship can significantly enhance a team's responsible AI maturity and ability to integrate responsible IA practices into their daily work.This involves not only raising awareness of RAI concepts but also developing and implementing action plans that embed these concepts into their existing work.

As someone deeply entrenched in the small business world, I understand the unique challenges faced by small and medium enterprises (SMEs) in the AI landscape: limited resources, lack of specialized expertise and primary focus on product development with little capacity for RAI practices. Our project has made significant strides in making SMEs overcome these limitations.

The project led to useful insights. For example, some teams confuse the ethics of privacy with data protection or GDPR, or will narrow down transparency to explainability. When we go over the review of the AI product, teams usually have a ''aha!'' moment when they find out that responsible AI requirements are not just an add-on to what they have developed, and that it needs to be factored in from the beginning, from their functional requirements. Further, they realise that scaling is not just dimensioning their solutions, but also verifying that they are well received worldwide and across cultures, and that scaling is also in responsible AI.

How do I envision implementing responsible AI? I think that it's first by creating a strong leading team in a company that puts the professional priority to define the conduct, and then devoting the resources to educate, implement, measure, and supervise seriously the scalability and the advancement of responsible AI. I believe in a multi-disciplinary approach to those challenges. Without a holistic thinking that takes into consideration various aspects, both technological, ethical, regulatory and soft skills, we will not be able to implement responsible AI.

Participating in the SRAIS mentorship was a transformative experience for our team because the program provided access to expert mentors (expertise), who helped us refine our approach to privacy, bias mitigation, and responsible AI development in the health care sector.

Sometimes it’s enough to ask the good question. GPAI experts ensured us that ethical AI is not a limitation but a competitive advantage. When AI is designed responsibly from the ground up (responsibility-by-design), there is no need for excessive regulatory intervention—because the technology itself aligns with the values we prioritize. This approach ensures that AI not only complies with regulations but remains valuable and beneficial to users, without compromising privacy or ethics.