The legal ground for processing the training data

8vance has developed a platform to connect people and work using advanced matching technology. By deploying smart AI, the 8vance platform can find the most suitable candidates, based on skills, personality, education, work experience and ambition. Factors like origin, gender, and age, which could lead to undesirable discrimination, are not included in the matching process. Bias, as often present in selection procedures, is therefore avoided as much as possible with the 8vance platform.

To develop and improve the platform, 8vance processed personal data to compile a training dataset in accordance with the General Data Protection Regulation (GDPR, or AVG in Dutch). Article 6 of the GDPR states that the processing of personal data by companies is allowed if there is a lawful basis for doing so.

How 8vance manages the processing of personal data is explained in detail in the privacy statement. This article goes further into the legal basis of the legitimate interest (Article 6(1) under f GDPR). Here, we provide a concise and clear explanation of the various interests weighed against each other to assess whether there is a legitimate interest in using the training data.

 

Composition Training data

The training data consists of publicly accessible profiles taken from various professional networking platforms on the internet. The information in the dataset consisted of the following fields:

First name, last name, contact information, employers, positions, geolocation (country and/or city), education, skills, experiences, dates corresponding to the employment period or course dates, link to original profile.

The data was anonymized to ensure that individuals could not be re-identified. This process included removing identifiable information such as names, contact details, company affiliations, and specific job titles (like prime minister). The use of embedding was included as well. Embedding is a way of storing or processing information without the original, data being visible or accessible. This involves converting the words or phrases containing sensitive information into an abstract numeric representation (an embedding) that cannot be traced back to the original words. This representation provides the key information for analysis while deliberately omitting any irrelevant data that could be sensitive or potentially used to identify an individual.

 

Finally, the original dataset containing all the data was deleted after processing. The remaining dataset thus consists of a combination of educational qualifications and job positions with associated experiences and skills that cannot be traced back to a specific individual. This means that the training data no longer contains personal data.

According to the Dutch Data Protection Authority’s guideline, three cumulative conditions must be met when assessing legitimate interest. In the following section, we outline these conditions while considering the relevant characteristics of training data as specified in the guideline.

 

Legitimate interest

8vance not only has a commercial interest but also serves a specific public interest. Namely, to help jobseekers find work that matches their knowledge, skills, abilities, and ambitions while simultaneously enabling employers to identify the most objectively qualified candidates, even in situations where the pool of suitable applicants appears to be depleted.

A legitimate interest involves interests where we in society attach importance to protection. The application of 8vance contributes to the protection of human rights, more specifically non-discrimination, by contributing to a transparent and accessible labour market. In a more transparent labour market, individuals can more easily identify where their opportunities exist. Many more people can thus participate in the labour market, proceed to do meaningful work, and become happier in their jobs.

Matching candidates based on their knowledge, skills, abilities, and ambitions can make the hiring process more objective and minimize the potential for discrimination. Only technology like the one from 8vance can determine whether (and how well) someone fits a job when someone does not have the standard education and work experience as normally expected. This is of great social importance, given the huge staff shortages in crucial sectors at the moment.

 

Necessity

To achieve matching based on objective criteria, the algorithm needs to be trained on large amounts of data. This is crucial to ensure that the data is diverse, representative, and up-to-date, to accurately represent a cross-section of the labour force. This data was collected from online profiles made public and kept up-to-date by the individuals themselves from different sectors and locations. Thus, to create a well-functioning algorithm, it was necessary to use this data. In this instance, no alternative, less intrusive methods are available to obtain the appropriate training data.

 

Balance of interests

Data subjects are the people whose personal data have been collected and processed from the public business profiles. Data subjects themselves have made their business profiles public through an online networking platform with (partly) the purpose of being findable for contact with potential new employers or clients. As the personal data shared were entered and made public for professional visibility, it is plausible that this data is used for finding career opportunities and matching vacancies. It is relevant to mention here that 8vance has no position of authority over the data subjects and cannot influence their behaviour in any way.

8vance has a significant interest in the optimal functioning of the matching tool, which is only possible with the help of this training data. At the same time, data subjects have a fundamental right to privacy that must be respected and protected at all times. When personal data is processed, data subjects lose control over their data, which poses risks such as identity theft or reputational damage. Ensuring this privacy is therefore crucial in this case because processing their personal data can damage their private lives. This risk arises particularly when data subjects do not want certain information about them to be shared with third parties.

8vance has taken measures to limit the risks and reduce the impact on the privacy of data subjects as much as possible. For example, 8vance has limited itself to processing only the strictly necessary personal data and anonymizing the data. In addition, 8vance implements comprehensive security measures that comply with the ISO 27001 standard. These measures aim to strictly limit access to the dataset, preventing unauthorized access.

This approach ensures that 8vance processes personal data responsibly (in line with the AVG), taking into account the rights of individuals, the interests of society and the operational needs of the organization.