Written by Anna Apergi Konstantinidis, Special Advisor Diversity Charter Greece
Examples from large companies worldwide have shown us that personnel selection algorithms rejected candidates based on social criteria. Not because someone programmed it that way, but because the data showed “what had worked in the past”. A typical case is that of a large and well-known multinational, which was forced to withdraw its AI CV evaluation system, as the algorithm “preferred” and simultaneously only recommended candidates who were cis* heterosexual, white men, due to the fact that it was based on biased, stereotypical historical data .
Algorithms, despite their apparent neutrality, reflect and often reinforce, if not complement, the social biases and stereotypes already present in their data. For example, a hiring algorithm trained on historical data that under-represents women and femininity in general, or people of non-Western ethnicities, LGBTQ+ people, or people with disabilities, is likely to internalize such biases and reproduce them.
However, Artificial Intelligence can become a powerful ally in favor of the inclusion of diversity, if it is designed with transparency, accountability, and ethical criteria.
Some examples of good practices include:
• Algorithmic Transparency and Bias Auditing : Methods such as “fairness testing” help detect and correct inequities in the output of AI systems.
• Anonymous CVs and “blinding” of sensitive characteristics: AI can contribute to the depersonalization of the evaluation of candidates, focusing on skills rather than social characteristics.
• Creating predictive models to identify exclusions: Patterns that indicate systemic exclusions can be identified, such as lower probabilities of progression for specific social groups, especially the most vulnerable.
In Greece, the Diversity Charter for Greek Businesses , Organizations and Institutions, an initiative of the European Commission , promotes equality, inclusion and diversity in the workplace. Its aim is to encourage Greek businesses to commit themselves ethically and operationally to creating inclusive environments.
The integration of Artificial Intelligence (AI) tools in Greek businesses that have signed the Charter can act as an accelerator of its purpose, provided that technological solutions should be designed with people, diversity and inclusion at the center. Especially in small and medium-sized businesses, where culture often depends on a few people, AI can offer a framework for more objectively evaluating and strengthening inclusion policies.
To ensure that AI works for diversity and not against it, I could suggest the following:
•Institutional integration of the Principles of Diversity at every stage of algorithm development.
• Training of all Technical Teams and Departments on Artificial Intelligence issues, combined with an inclusive design.
•Mandatory transparency statement for AI systems that affect human lives.
•Continuous connection with European initiatives such as the Diversity Charter.
We will have the opportunity to say much more about Artificial Intelligence in Greek businesses at the 7th Diversity in Business Conference coming up in October, but what we must emphasize is that Artificial Intelligence is not an ally or an obstacle in itself. It is a valuable tool and we should treat it as such. The direction and evolution it will take depends on the intentions, values, and responsibility of those who design and use it.
In an era when diversity is a fundamental condition for sustainability and innovation, the combination of technology and the human code of values, human and individual rights, can make Artificial Intelligence one of the greatest and most powerful tools for a fairer and more inclusive world.
Terminology:
* 4th Industrial Revolution: The current trend of automation and data exchange in manufacturing technologies has been called the Fourth Industrial Revolution. It includes cyber-physical systems, the Internet of Things, cloud computing, and cognitive computing.
* Artificial Intelligence: The term Artificial Intelligence (AI) refers to the field of computer science that deals with the design and implementation of computer systems that mimic elements of human behavior that imply even rudimentary intelligence: learning, adaptability, inference, contextual understanding, problem solving, etc.
* CIS: The term “cis” or “cisgender” refers to individuals whose gender identity (social gender) does not differ from the sex assigned to them at birth (biological/anatomical sex).
