Mitigate ChatGPT Privacy Concerns with PrivateGPT Headless

It’s as simple as a few lines of code around each OpenAI call. Data is de-identified and then re-identified by a container running in your or your customer’s premises. No data is ever shared with Private AI. 

				
					import openai
from privateai import PrivateGPT

MODEL = "gpt-3.5-turbo"
messages = [{"role": "system", "content": "You are an email answering assistant"},
            {"role": "user", "content": "Invite Keanu Reeves for an interview on April 19th"}]

privategpt_output = PrivateGPT.deidentify(messages, MODEL)
response_deidentified = openai.ChatCompletion.create(model=MODEL, messages=privategpt_output.deidentified, temperature=0)
response = PrivateGPT.reidentify(response_deidentified, privategpt_output)

				
			

Request an API key to go Headless today or test it yourself for free with the PrivateGPT Chatbot demo

The Problem:

There's a wave of ChatGPT-powered applications, but enterprises don’t want their data sent to ChatGPT

Generative AI, such as OpenAI’s ChatGPT, has enabled a range of new applications. In particular, many ML-powered applications such as summarisation and chatbots are moving to ChatGPT. 

The problem is that most enterprises have blocked ChatGPT internally and aren’t OK with their data being sent out of their systems to OpenAI. And for good reason too, ChatGPT was temporarily banned in Italy and has already had its first data leak, which exposed personal information including some credit card details. 

Enterprises also don’t want their data retained for model improvement or performance monitoring. This is because these systems can learn and regurgitate PII that was included in the training data, like this Korean lovebot started doing, leading to the unintentional disclosure of personal information. 

Enter PrivateGPT:

Easily identify and remove 50+ types of PII inside your application before sending it through to ChatGPT

With the help of PrivateGPT, developers can easily scrub out any personal information that would pose a privacy risk, and unlock deals blocked by companies not wanting to use ChatGPT.  

With PrivateGPT Headless you can:

laptop with a chat log on screen and a lock in front

Why Private AI

Deploys as a Docker container - no data is ever shared with us

50+ entity types covering all major regulations like GDPR, HIPAA & PCI DSS

Runs in real-time - the user won’t notice it’s there

Built for scale. A single container can service thousands of requests/sec

Advanced re-identification to put PII back into the response from OpenAI

Works with any LLM service, such as Anthropic and Cohere

We understand the significance of safeguarding the sensitive information of our customers. With Private AI, we can build our platform for automating go-to-market functions on a bedrock of trust and integrity, while proving to our stakeholders that using valuable data while still maintaining privacy is possible.

Sunil Rao
CEO, Tribble

Download the Free Report

Request an API Key

Fill out the form below and we’ll send you a free API key for 500 calls (approx. 50k words). No commitment, no credit card required!

Language Packs

Expand the categories below to see which languages are included within each language pack.
Note: English capabilities are automatically included within the Enterprise pricing tier. 

French
Spanish
Portuguese

Arabic
Hebrew
Persian (Farsi)
Swahili

French
German
Italian
Portuguese
Russian
Spanish
Ukrainian
Belarusian
Bulgarian
Catalan
Croatian
Czech
Danish
Dutch
Estonian
Finnish
Greek
Hungarian
Icelandic
Latvian
Lithuanian
Luxembourgish
Polish
Romanian
Slovak
Slovenian
Swedish
Turkish

Hindi
Korean
Tagalog
Bengali
Burmese
Indonesian
Khmer
Japanese
Malay
Moldovan
Norwegian (Bokmål)
Punjabi
Tamil
Thai
Vietnamese
Mandarin (simplified)

Arabic
Belarusian
Bengali
Bulgarian
Burmese
Catalan
Croatian
Czech
Danish
Dutch
Estonian
Finnish
French
German
Greek
Hebrew
Hindi
Hungarian
Icelandic
Indonesian
Italian
Japanese
Khmer
Korean
Latvian
Lithuanian
Luxembourgish
Malay
Mandarin (simplified)
Moldovan
Norwegian (Bokmål)
Persian (Farsi)
Polish
Portuguese
Punjabi
Romanian
Russian
Slovak
Slovenian
Spanish
Swahili
Swedish
Tagalog
Tamil
Thai
Turkish
Ukrainian
Vietnamese

Rappel

Testé sur un ensemble de données composé de données conversationnelles désordonnées contenant des informations de santé sensibles. Téléchargez notre livre blanc pour plus de détails, ainsi que nos performances en termes d’exactitude et de score F1, ou contactez-nous pour obtenir une copie du code d’évaluation.

99.5%+ Accuracy

Number quoted is the number of PII words missed as a fraction of total number of words. Computed on a 268 thousand word internal test dataset, comprising data from over 50 different sources, including web scrapes, emails and ASR transcripts.

Please contact us for a copy of the code used to compute these metrics, try it yourself here, or download our whitepaper.