Policy Statement on Compliance with the EU AI Act

### Introduction

At Ferretly, we are committed to providing our clients with insightful reports on candidate behaviors sourced from publicly available social media data while adhering to the European Union’s Artificial Intelligence Act. This policy outlines how Ferretly describes ensuring compliance with that regulation, available at Regulation (EU) 2024/1689.

### 1. Human–AI collaboration

AI as a tool: Ferretly’s service utilizes AI internally to assist human analysts in identifying and reporting on specific behaviors from social media. Per the published policy, the AI does not make decisions or generate scores for that decision purpose; it aids in data processing to provide factual insights.

### 2. No automated decision-making or scoring (as described)

Human oversight: All analyses are overseen by human analysts who review AI-assisted findings. The published policy states this ensures reports do not contribute to or facilitate social scoring or automated decision-making prohibited under the EU AI Act.

### 3. Data governance and bias mitigation

Fairness in data use: Ferretly states that its AI tools are trained on data sets designed to minimize bias, and that regular audits are conducted to detect and correct potential biases, aligning with the Act’s requirements for high-risk AI (as described in the policy).

### 4. Consent and transparency

Informed consent: Ferretly requires explicit consent from candidates before analyzing their public social media profiles. The consent process informs candidates how their data will be used, emphasizing transparency in line with both the EU AI Act and GDPR principles (as stated in the policy).

Communication of use: Ferretly states it clearly communicates to both candidates and employers that the service provides informational reports without influencing decisions through scoring or profiling.

### 5. Handling of public data

Respect for privacy: Although social media data is public, Ferretly states it handles that data with respect for individual rights, and that its approach aims to comply with the EU AI Act’s ethical considerations regarding data usage even when data is publicly accessible.

### 6. Reporting to employers

Informational reports: Reports are structured to provide factual insights about candidate behaviors, avoiding implications that could lead employers to engage in social scoring or discriminatory practices. Emphasis is on information, not on decision-making.

### 7. Compliance documentation

Record keeping: Ferretly states it maintains comprehensive records of compliance efforts, including how consent is managed, how data is processed, and how human oversight is integrated into service delivery.

### 8. Continuous compliance monitoring

Adaptability: As interpretation of the EU AI Act evolves, Ferretly commits to updating practices for ongoing compliance, including regular compliance checks and updates to policies, procedures, and AI tools.

### 9. Legal and ethical compliance

Expert consultation: Ferretly states it engages legal experts specialized in EU law to ensure interpretation and implementation of the EU AI Act’s requirements remain correct and up to date.

### Conclusion

Ferretly’s commitment to compliance with the EU AI Act is described as integral to its operations—by combining AI technology with human oversight, ensuring transparency, and focusing on ethical data use—to provide insights while safeguarding the rights and dignity of individuals involved.

### Related articles