andreas.lopez93andreas.lopez93
andreas.lopez93
Oct 22, 2023

Q: The documents uploaded and questions+ answers queried - will they be 100% private or will they somehow end up in a cloud or such?

I.e. chatGPT has the issue where people were able to query it and get answers for other people's private work and questions.

We are working with very confidential documents and it would be great to have a tool like this, but we are worried about data leaks like what happened to Samsung.

Founder Team
alecngn

alecngn

Jun 27, 2024

A: Happy weekend Andreas,

Thank you for the question and that's a very valid concern I'm happy to address.

It’s important to consider data security practices when using any AI research tool. My team and I strive to provide a secure and reliable platform, but it’s always advisable to exercise caution. Here's more information on our security:
https://help.afforai.com/en/articles/8478761-how-secure-is-my-data-with-afforai

All articles and and files you uploaded to Afforai will be stored on the Azure Enterprise Cloud until you decide to delete them. The files are encrypted and we NEVER have access to your documents. Here's more info on our data security practices: https://afforai.com/security

You can connect any of your files to your chatbots to train, extract, summarize, translate stories, texts. If the file is removed from the file storage, the knowledge will be deleted as well. The answers will always be there until you clear the conversation with the chatbot.

I hope I answered all of your questions. I'm here ready to assist further. Thank you for your support.

Share
Helpful?
2
Log in to join the conversation

Excellent question for a research tool that might include a researchers proprietory intellectual information. We have cases of corporate espionage with Countries stealing proprietory information all the time now.
How safe is the information, you mention encryption, but how can you secure information that is fed into ChatGPT as it is surely storing the query information that is submitted to the AI model for processing? So in effect any documentation loaded into an AI model is stored by the AI provider and can be data mined at their leisure? How can proprietory IP information ever be loaded safely into an AI tool?

I just popped this into an AI tool to see what answer it would giveThe ability for an AI model owner to data mine proprietary IP information loaded into their software depends largely on the terms of service and privacy policies agreed upon between the parties involved. However, ethical and legal norms generally prohibit such behavior.

Consent: The use of proprietary IP information should be based on explicit and informed consent of the owner. If the owner has not granted permission for their data to be mined, it should not be.

Privacy Policies: The AI software must abide by its stated privacy policy. If this policy states that data will not be mined or used for purposes other than those explicitly agreed upon, then it should not happen.

Legal Compliance: There are numerous laws and regulations in place to protect intellectual property and personal data. Breaching these can result in serious penalties.

Encryption and Anonymization: If proprietary IP information is encrypted or anonymized, it becomes more difficult for it to be mined in a meaningful way. However, this does not completely eliminate the possibility.

Access Control: Strict access control protocols can reduce the risk of unwanted data mining by limiting who has access to the data.

If you're concerned about data mining, it's crucial to have a clear understanding of the terms of service and privacy policies of the AI software you're using, and consider seeking legal advice if necessary.

Founder

Hi ManyWays,

Thank you for the question and that's a very valid concern I'm happy to address.

It’s important to consider data security practices when using any AI research tool. My team and I strive to provide a secure and reliable platform, but it’s always advisable to exercise caution. Here's more information on our security:
https://help.afforai.com/en/articles/8478761-how-secure-is-my-data-with-afforai

But to also give you an answer here, your concern is the exact reason why we choose to use Microsoft Azure OpenAI API instead of the public OpenAI API.

All articles and and files you uploaded to Afforai will be stored on the Azure Enterprise Cloud until you decide to delete them. The files are encrypted and we NEVER have access to your documents. Here's more info on our data security practices: https://afforai.com/security

You can connect any of your files to your chatbots to train, extract, summarize, translate stories, texts. If the file is removed from the file storage, the knowledge will be deleted as well. The answers will always be there until you clear the conversation with the chatbot.

At the end of the day, it's up to you to decide whether our security measures is to the level that you'd be okay with so my team and I can only try our best to create a safe, secure, and reliable research tool and I think our efforts reflect that.

I hope I answered all of your questions. I'm here ready to assist further. Thank you for your support.