1) Introduction and Commitment
Philologica is committed to the responsible, ethical, and transparent development and use of Artificial Intelligence. We believe AI should augment, not replace, human judgment, particularly in the academic field. This policy outlines our approach to complying with the EU Artificial Intelligence Act (Regulation (EU) 2024/1689) and ensuring our service is trustworthy and secure for all users.
2) Risk Classification
Philologica is designed to support the academic review process by providing structured analysis, suggestions, and reports. Our service is not intended to make automated decisions. The final evaluation and editorial judgment always rest with the human user.
Based on this design, Philologica is classified as a minimal or no-risk AI system under the EU AI Act. We do not engage in prohibited practices nor fall under the category of high-risk systems, as we do not, for example, perform automated grading, determine access to educational programs, or evaluate individuals in a way that affects their legal status or fundamental rights.
3) Core Principles
Our AI policy is built on the following core principles:
- Human Oversight: We believe in the essential role of human review. All our reports are designed as decision-support tools. We explicitly state that "Human review is always recommended" and that final academic judgment remains with the user.
- Transparency: We are open about the use of AI in our processes. We are committed to clearly identifying AI-generated content within our platform, in line with the transparency obligations of the AI Act.
- Privacy and Data Protection: The security and privacy of your documents are paramount. We adhere to strict data encryption and retention policies. Your submitted content is not used to train our models, ensuring the confidentiality of your work.
- Security: We are committed to maintaining a high level of robustness, cybersecurity, and accuracy in our systems to protect our users' data and the integrity of our service.
4) Transparency and Disclosure
In accordance with Article 50 of the EU AI Act, we have implemented the following transparency measures:
- Direct Interaction: Users are informed when they are interacting with an AI system within the Philologica workspace.
- AI-Generated Content: All content generated by our AI, such as findings, recommendations, and executive summaries within a report, will be identifiable as AI-generated. We are actively implementing measures to mark this content, and we will adopt the EU's recognized icons and machine-readable marking solutions as they become available. Our AI performs an assistive function for standard editing and does not substantially alter the input data without human review, which qualifies for certain exemptions under the guidelines, but we will mark content proactively.
- Deepfakes and Public Interest Text: As Philologica does not generate manipulated image, audio, or video content, the requirements for deepfake labeling are not applicable to our core service. If we were to publish text to inform the public on matters of public interest, it would undergo human review and editorial control before publication, thus not requiring an AI-generated label.
5) AI Literacy
We are dedicated to fostering a sufficient level of AI literacy among our staff and users. In line with Article 4 of the EU AI Act, we will:
- Provide clear and understandable information about how our AI tools work, their capabilities, and their limitations.
- Offer guidance and resources to help our users make informed decisions based on our reports.
- Ensure our internal team has the necessary skills and understanding to develop, operate, and deploy our AI systems responsibly.
6) Intellectual Property and Copyright
We respect intellectual property rights. Our AI processes are designed to analyze academic texts and verify sources against public academic databases (e.g., Crossref, OpenAlex). We operate in accordance with EU copyright law and are committed to using state-of-the-art technologies to identify and comply with any reservations of rights regarding the content used in our training data.
7) General-Purpose AI Models
We may utilize certain third-party general-purpose AI models (GPAI) as part of our processing pipeline. We ensure that the providers of these models are compliant with the EU AI Act's requirements regarding technical documentation, copyright policies, and transparency. We cooperate with the providers to obtain the necessary information for our compliance and to provide appropriate documentation to our users.
8) Accountability and Contact
We take our responsibility as a provider of AI systems seriously. We have implemented internal compliance procedures and remain available to answer any questions you may have regarding this policy or the use of AI in our services. For any inquiries, please contact us through the channels provided on our website ([email protected]).
9) Policy Review
This policy will be reviewed and updated regularly to reflect changes in our services, user feedback, and the evolving legal and regulatory landscape for AI, such as the implementation of the AI Act's requirements.