On January 17th, ALFIE held an internal workshop with project partners. Led by the Universitat Autònoma de Barcelona (UAB), the workshop aimed to organise and categorise the user requirements gathered during the initial stages of the project. Using the interactive platform Miro, partners sorted the user requirement gathered by UAB into functional and non-functional requirements. Here are some of the results from the workshop: 

Defining Key Terms:

Functional Requirements: These define the specific capabilities and behaviours expected from the AutoML platform. Essentially, they outline what the system must do. Examples from the workshop include:

• • Ensure unbiased recognition of accents, dialects, and speech impairments.
  • Ensure AI systems support both oral and written communication in simple, clear language. 
  • Users should be able to verify the trustworthiness and credibility of AI outputs. 

Non-Functional Requirements: These focus on the quality attributes and constraints that govern the system’s performance and usability. They address how the system should function. Examples from the workshop include:

• • • AI systems should be energy efficient and minimise environmental impact. 
    • Ensure that AI systems are explainable and guide users through decision-making processes. 
    • AI outputs (e.g. images or responses) should reflect diversity in gender and ethnicity, and cultural contexts.

Users 

We also discussed the potential user groups for the ALFIE AutoML platform and divided them according to their unique requirements.

• Designer Users: These individuals would leverage the platform to construct ethical AI tools. This category encompasses a diverse range of professionals, including designers, developers, and even those with limited technical expertise who aim to build AI solutions tailored to their specific needs.
  
  
• End Users: This group comprises individuals who would directly interact with the AI models generated using the AutoML platform.
• Workshop Collaboration

The workshop was a collaborative effort, with contributions from all project partners, including the leaders of the three use cases: UAB, Catalink, and Bosch. Future plans for Use Cases and defined terms for the project moving forwards were also discussed. 

Next Steps: Technical Requirements

The valuable insights and requirements from this workshop will serve as the foundation for drafting the technical requirements for the AutoML platform. This phase will be undertaken by project partners at the University of Eindhoven. The technical requirements will provide a detailed roadmap for the platform’s development, ensuring that it meets the diverse needs and expectations of its intended users.

Conclusion:

The successful completion of this workshop marks a significant milestone in the ALFIE project. By organising and categorising user requirements, the project team has laid a strong groundwork for the development of a robust and user-centric AutoML platform. The collaborative spirit demonstrated by all partners bodes well for the project’s future success.