Is Data Input into Teachme AI Used for Retraining AI Models?
No, customer data is not used for retraining AI models.
Teachme Biz adopts a system that does not expose data externally.
How Are Uploaded PDFs, Videos, and Texts Stored?
The uploaded data itself is not stored.
Videos and texts registered in steps are stored for use as manuals in the customer environment.
Intermediate outputs (OCR results and LLM responses) are stored for troubleshooting and accuracy improvement purposes.
* Data storage methods and retention periods may change with service feature expansions.
Important changes will be announced separately.
Is There Any Risk of Data Being Mixed with Other Companies' Data?
No, in Teachme Biz, data is separated by company within Teachme Biz and cannot be mixed with other companies' data.
Who Owns the Copyright of Generated Manuals?
The copyright of generated manuals belongs to the customer.
Teachme AI functions as a tool to assist in content generation based on customer input.
For details, please refer to Article 18 "Protection of Contractor Content" in the Teachme Biz Terms of Service.
Related: Teachme Biz Terms of Use
https://file.biz.teachme.jp/doc/teachmebiz_terms_e.pdf
How is the Quality of AI-Generated Results Ensured?
For each type of AI feature, verification is conducted to ensure intended generation results are achieved with certain probability or accuracy levels.
However, due to AI characteristics, generation results cannot be completely controlled.
Teachme AI is designed to allow confirmation and editing of generation results before publication.
Final quality control should be performed by the customer.
How Are Inappropriate AI Generations Handled?
In the unlikely event of inappropriate generation, immediate correction or deletion is possible.
Reports to customer support will be used to improve the system.
Related: Contact Form
https://forms.gle/t6LbL2MhY155qptn9
Which Generative AI Models Are Being Used?
As of December 2, 2024, multiple Teachme AI features use Anthropic's Claude model, and some features combine models from companies like Amazon.
The system configuration is continuously improved to provide appropriate performance to customers by reviewing AI services and models used in line with AI technology evolution.
Model selection is comprehensively determined based on performance, proven enterprise use cases, and high security reliability.
Additionally, input data is not used for model retraining.