Titulo Estágio
Towards a Responsible AI Program in the Recruitment Domain
Áreas de especialidade
Sistemas de Informação
Sistemas Inteligentes
Local do Estágio
DEI-UC / skeeled
Enquadramento
The European Union recently launched the Coordinated Plan on Artificial Intelligence (AI), aiming to lead trustworthy AI adoption in organizations (European Commission, 2021). Therefore, IT companies adopting AI in their products will need to rethink their (1) software development lifecycles, (2) tools, (3) operation conditions, and (4) auditability procedures. Aspects like ethics in AI, sustainable, and trustworthy AI will be a priority in the agenda of IT companies. Ethics guidelines are already in place and the “era of ‘move fast and break things’ is over” (Floridi, 2019). Nevertheless, it is also expected that current recommendations and suggestions may be turned into legal obligations or standards in the near future. Ultimately, the capacity to adopt ethical and trustworthy AI may be a requirement to compete in the market.
Skeeled is an expert in talent acquisition, providing an all-in-one, AI-based software that streamlines the companies’ hiring process, including profile screening, personality evaluation, video interviews, and candidate ranking, with headquarters in Luxembourg and development team in Porto's office. Working for major European companies in several industries (e.g., automotive, healthcare, and air transportation), our products use advanced data analytics and AI techniques to support hiring decisions. Trustworthy AI is an important pillar of our strategy to assist companies in making the best decisions, in a fair and transparent way. Therefore, we are preparing our software/models development team to the upcoming regulations in trustworthy AI.
The candidate will have the opportunity to work with our office in Porto in one of the most challenging areas for AI projects: ensure that the company adopts the best practices in their development and decision support processes and is able to provide evidence of compliance to third party entities. The internship will involve a multidisciplinary approach to AI adoption, including topics related to Information Systems, Software Engineering, and Intelligent Systems. Candidates will have the opportunity to develop skills in information systems management, IT audit, continuous delivery and automation pipelines, and trustworthy AI adoption in decision support systems.
References:
European Commission. (2021). Coordinated Plan on Artificial Intelligence 2021 Review. Retrieved from https://digital-strategy.ec.europa.eu/en/library/coordinated-plan-artificial-intelligence-2021-review
Floridi, L. (2019). Establishing the rules for building trustworthy AI. Nature Machine Intelligence, 1(6), 261–262. https://doi.org/10.1038/s42256-019-0055-y
Objetivo
The main objectives of this work are:
1. Evaluate the requirements of trustworthy AI adoption in intelligent recruiting platforms;
2. Propose a comprehensive transformation plan for skeeled business. This multi-year plan should incorporate recommendations for the tools used in the software development, datasets, models lifecycle, control objectives for AI adoption during development and operation stages;
3. Develop a web-based tool to support the audit of trustworthy AI adoption at skeeled.
Plano de Trabalhos - Semestre 1
The main phases for the first semester are presented as follows:
1) Study of the state-of-the-art: EU regulations on ethics and trustworthy AI. Scientific literature review
Start: September 2021, Duration: 1 month
2) Evaluate the most relevant requirements for trustworthy AI adoption at skeeled
Start: October 2021, Duration: 2 month
3) Draft the requirements for the supporting tool (technologies to be selected by the candidate).
Start: November 2021, Duration: 1 month
Results of the 1st semester:
- Understand the importance of trustworthy AI for the future of IT companies
- Architecture for a tool that supports internal audits at skeeled.
Intermediate report.
Plano de Trabalhos - Semestre 2
The main phases for the second semester are presented as follows:
4) Create a strategic plan for trustworthy AI at skeeled. The plan will include measures for the development team (e.g., how to document datasets of recruitment candidates and to compare the tools used in the development process and its suitability to prove trustworthy AI), for the tools used in the development (e.g., suitability to prove trustworthy AI, search for alternative tools that could be implemented at skeeled), and in the support of our customers decisions (e.g., evidences that need to be provided about ethics and transparency in the recruiting suggestions, including those provenient from distillation analysis of the used models)
Start: January 2022, Duration: 4 months
5) Develop a supporting tool (prototype) for auditing trustworthy AI. The tool is complementary to the main goal of the internship and has two main purposes: (1) provide a centralized database for audit reports and (2) provide evidence to third parties about compliance to trustworthy AI practices.
Start: February 2022, Duration: 3 months
6) Writing of the thesis and a scientific paper.
Start: May 2022, Duration: 2 months
Results of the 2nd semester:
• Final report
Condições
The student will be integrated in the Information Systems group (ISG) of CISUC, Department of Informatics Engineering, University of Coimbra. A workplace and the required resources will be provided.
The work will be mostly remote. Travels to Porto, if necessary, are supported by skeeled.
Observações
This work would be supervised by UC-DEI Professors João Barata (supervisor) Catarina Silva (co-supervisor).
Orientador
Joana Urbano
joana.urbano@skeeled.com 📩