AI Safety Briefing
Understanding AI Hallucinations
One of the inherent challenges in all current AI technologies, including those used in Socrat.ai, is the risk of generating inaccurate or misleading information, commonly referred to as "hallucinations." This can occur even with advanced models, as the AI attempts to generate helpful responses based on the patterns it has learned from the training data. While we continuously work to improve the accuracy of Socrat.ai, we acknowledge that it may occasionally produce non factual or unintended content. We encourage educators to explain these risks with their students, and for all parties to critically evaluate AI-generated information, utilizing it as a complement to traditional teacher-led learning.
Commitment to Data Privacy
Protecting user privacy and securing personal information are paramount at Socrat.ai. We adhere to strict data protection protocols to ensure that all user data is handled with the highest degree of security. We do not sell any user data nor do we use personal data to train our or other AI models. Our AI is trained on a broad and diverse set of data sources that are completely unrelated to our users' personal data, ensuring that the integrity and privacy of user information are maintained at all times.
Tools for Educators
We understand the importance of transparency and oversight in educational environments. Socrat.ai provides educators with the ability to monitor student learning and progress. Our tools, including Socrat Play, help educators ensure that the AI is used appropriately and effectively, enhancing the learning experience while safeguarding the educational process.
Data Retention and Deletion Policy
Socrat.ai is committed to maintaining a high level of data integrity and security. All user data is securely stored and is automatically deleted from our servers after 30 days. This policy helps ensure that user data does not persist indefinitely on our platforms and is aligned with best practices for data management and compliance with global data protection regulations.
EdSafe AI Industry Council
We are a proud member of the EdSafe AI Industry Council, which is comprised of "leading education and technology companies of all sizes committed to ensuring artificial intelligence and related technologies – in their own companies’ solutions and across the education landscape – are safe, accountable, fair, and effective (SAFE). " We pledge to create and promote responsible AI technologies and practices, and work to consistently demonstrate and uphold this commitment.
Our Commitment to Schools
At Socrat.ai, we are willing to sign individualized policies in order to demonstrate our commitment our client institutions. As we continue to advance in our AI capabilities, we remain vigilant about the ethical implications and potential risks. Our dedication to improving Socrat.ai responsibly is driven by our core values of trust, transparency, and user safety. We are committed to delivering a reliable and secure AI tool that supports educational goals and fosters a safe and responsible learning environment. By openly discussing the potential limitations and challenges of AI, and by implementing rigorous safety and privacy measures, we aim to build trust with our users and create a more informed and cautious approach to AI interaction.
Compliance with AI in Education Regulations
Socrat.ai ensures the privacy of student education records and operates in full compliance with FERPA. All student data is under the direct control of the educational institution, and Socrat.ai strictly adheres to parental consent requirements for access to children’s educational records.
Socrat.ai requires parental consent for users under the age of 13, as mandated by COPPA. We do not collect more personal information from children than is necessary, ensuring that all data handling practices are age-appropriate and secure.
