This guide includes steps to limit the risks of using this tool such as:
Enabling Two-Factor authentication
Managing if and how your chats are shared
Reviewing third party apps that ChatGPT can access
Preventing use of your data to train the model
Deleting chats
Turning Off Ad Personalisation

Enable Two-Factor authentication
Two-Factor authentication provides an extra security step in order to access your account. This way, besides asking for your username and password, ChatGPT will protect your account access by requiring a unique code on your device. This reduces the chances of someone accessing your account and getting access to your previous chats.
You will need a dedicated app to setup Two-Factor Authentication.

On the mobile app:
Go to your profile by clicking the menu in the top left corner and tapping on your profile in the bottom left corner. Navigate to Security and tap multi-factor verification.
A browser window will open and you will be asked to login. Follow the instructions to finish the setup.

On your browser:
Go to the settings by clicking on your profile in the bottom left corner and clicking Settings. Navigate to Security and click multi-factor verification.
You will be redirected to a new page, follow the instructions to finish the setup.

Managing if and how your chats are shared
ChatGPT allows you to share conversations via a URL. While in theory only people with the link can read the conversation, it’s not impossible for this link to end up online and be indexed by search engines. Before sharing a link to a conversation, you should ensure it does not contain personal or sensitive information that you don’t want to share.
Note that this feature is only available on your browser and not in app*

You can manage your previously shared chats in the settings by clicking on your profile in the bottom left corner and clicking Data Controls. Click Manage next to Shared links and delete any chat you don’t want to be public anymore.

Review third party apps that ChatGPT can access
ChatGPT can access documents, calendar events and other information from connected apps, such as cloud file storage (e.g. Google Drive). We have expressed concerns about the privacy and security risks created by connecting AI chatbots to other apps as it can be exploited to access personal data.

To limit these risks, you should regularly review which apps ChatGPT is connected to and remove services you don’t or no longer use. This is particularly important if you are using ChatGPT Plus or Pro as more connected apps are available and ChatGPT agent mode can use those apps with limited control from the user.

Note that this feature is only available on your browser and not in app

Go to the settings by clicking on your profile in the bottom left corner and clicking Settings. Navigate to Connected apps and review the apps you are connected to.

Once this is done, go through the settings of the app you had previously connected and remove ChatGPT from the third party apps with access.

Prevent use of your data to train the model
Large Language Models are trained on massive datasets. We have expressed concerns about the lawfulness of using data collected online for this purpose and there are demonstrable risks of this data being reproduced by AI chatbots without user consent or knowledge (called “regurgitation”). To limit the risk of personal data leaking out, we suggest preventing the content of your chats from being used for further training.

On the mobile app:
Go to your profile by clicking the menu in the top left corner and tapping on your profile in the bottom left corner. Navigate to Data Controls and turn off Improve the model for everyone.

On your browser:
Go to the settings by clicking on your profile in the bottom left corner and clicking Settings. Navigate to Data Controls and turn off Improve the model for everyone.

Delete chats
Chats, like emails or messages, might contain sensitive information you don’t want others to access. To limit the risks in case of unauthorised access by someone with access to your device or a malicious actor, it can be good practice to delete chats you won’t come back to.

On the mobile app:
Tap on a chat and click the three dots in the top right corner. Tap delete.

Delete a conversation on mobile
On your browser:
Click the three dots next to a chat in the left-hand bar and click delete.
You can also delete all your chats in the Data Controls menu.

Turn Off Ad Personalisation
ChatGPT has started to introduce ads in some regions, along with controls that let you limit how much your activity shapes what you see. For now this feature is only available to users in the United States as part of a limited rollout and there is a setting to turn off ad personalisation and reduce how much of your data is fed into this targeting.

This feature will be coming to other regions, so it’s worth checking your settings regularly. When it appears, you’ll find it under Settings > Ad Controls.

If you turn ad personalisation off you will still see ads, but based on the context of your current chat thread and not based on accumulated data from past conversations.

Links

• More about Edgar: https://www.lse.ac.uk/people/edgar-whitley

• More about the London School of Economics and ID: https://www.lse.ac.uk/research/research-impact-case-studies/scrapping-costly-controversial-proposals-identity-cards

• Gus and Edgar's joint publications:
  - Global challenges for identity policies: https://books.google.co.uk/books/about/Global_Challenges_for_Identity_Policies.html?id=P7hsCwAAQBAJ&redir_esc=y
  - Global Identity Policies and Technology:
  Do we Understand the Question? An open access paper: https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1758-5899.2010.00028.x
  - Book chapter on Aadhaar: https://researchonline.lse.ac.uk/id/eprint/90490/

• Edgar's report on Women and ID: https://silkstart.s3.amazonaws.com/1967006c-a430-414e-9982-75e204478d6e.pdf

• More information on Aadhaar in India: https://privacyinternational.org/news-analysis/819/aadhaar-scheme-faq

• The Indian Supreme Court decision: https://privacyinternational.org/long-read/2299/initial-analysis-indian-supreme-court-decision-aadhaar and https://www.scobserver.in/cases/beghar-foundation-ks-puttaswamy-aadhaar-review-case-background/

• The Windrush scandal: https://jcwi.org.uk/reportsbriefings/windrush-scandal-explained/

• Interviews with people in Kenya who had been double registered: https://media.privacyinternational.org/my-library/video-playlists/kysrYXu54t5ennt5NtjNxc

• Yes Minister TV show: https://archive.org/details/yes-minister-1980-1984 including the famous Big Brother episode: https://archive.org/details/yes-minister-1980-1984/Yes+Minister+-+S01E04+-+Big+Brother.mp4

• The Clarkson toothpick incident: https://pressgazette.co.uk/publishers/nationals/jeremy-clarkson-victim-of-id-fraud-after-publishing-bank-details/

• The Washington Post raid: https://theintercept.com/2026/01/30/washington-post-hannah-natanson-fbi-biometrics-unlock-phone/

Previous podcasts:

• Double Registration - The trouble with Identity in Kenya: https://media.privacyinternational.org/w/midJFm7fJiH8mfQMbY3gY9

Links:

• https://jamesmuldoon.org/
• https://www.faber.co.uk/product/9780571399277-love-machines/

Links

ACLU report: ⁠https://www.aclu.org/campaigns-initiatives/aclu-vs-trump⁠

Kyllo case: ⁠https://en.wikipedia.org/wiki/Kyllo_v._United_States⁠Underground

Empire book: ⁠https://henryfarrell.net/underground-empire-2/

⁠Swift work between PI and ACLU: ⁠https://privacyinternational.org/blog/1355/pi-and-aclu-show-swift-auditor-has-extensive-ties-us-government

Yale lectures on National Security law: ⁠https://epe.yale.edu/videos/above-law-or-beneath-it-how-national-security-law-subverts-accountability-and-weakens-0

Asylum for White South Africans: h⁠ttps://www.nytimes.com/2025/10/30/us/politics/trump-refugee-admissions-white-south-africans.html

FISA reauth: ⁠https://theintercept.com/2026/01/29/nsa-702-fisa-surveillance/

Privacy advocates from Big Brother Watch, Silkie Carlo, and Privacy International, Nuno Guerreiro de Sousa, have issued pre-action correspondence to the Home Office and Metropolitan Police and are calling for a moratorium on facial recognition searches of the passport and immigration databases.

Learn more by going to https://privacyinternational.org/news-analysis/5635/revealed-skyrocketing-scale-uk-polices-secret-facial-recognition-searches

In the UK, asylum seekers are detained - often for a long time - while their claims are processed. If they’re not deported, they get released, but many have to wear an ankle tag which the Home Office can use to follow their movements 24 hours a day.

But there don’t seem to be any rules about who does and doesn’t get tagged, and how the data these tags collect is used by the Home Office.

So we bought some tags from the open market and tested them, to see how they work.
#dignitynotdata #privacy #surveillance

Migrants in the UK are forced by the Home Office to wear GPS tracking devices - humiliated, controlled, spied on, stigmatised - even if they’ve lived in the UK for years.

Capita PLC is one of the companies outsourced by the UK Home Office to deliver this inhumane policy - earning £38 million a year for it.

They’re profiting from the hostile environment.

We’re appalled at this treatment. With BID and MO, we’re calling on Capita PLC to stop providing services to the Home Office.

#dignitynotdata #privacy #surveillance #gpstracker

Migrants in the UK are forced by the Home Office to wear GPS tracking devices - humiliated, controlled, spied on, stigmatised - even if they’ve lived in the UK for years.

Capita PLC is one of the companies outsourced by the UK Home Office to deliver this inhumane policy - earning £38 million a year for it.

They’re profiting from the hostile environment.

We’re appalled at this treatment. With BID and MO, we’re calling on Capita PLC to stop providing services to the Home Office.

#dignitynotdata #privacy #surveillance #gpstracker

“A privacidade é um tema muito importante para mim porque ela é um pilar que garante outros direitos”

We’re celebrating the amazing women who are fighting to protect privacy and create a better world.

Get inspired ✨ #IWD2023

“I became acquainted on the issues of privacy when working on the protection of undocumented migrants in Europe”

We’re celebrating the amazing women who are fighting to protect privacy and create a better world.

Get inspired ✨ #IWD2023 #Shorts

Read more of ADC's work in their report "Quién revisa tu teléfono?' part one available here: https://adc.org.ar/wp-content/uploads/2022/01/ADC-Quien-revisa-tu-telefono.pdf and part two available here: https://adc.org.ar/wp-content/uploads/2022/11/ADC-Quien-revisa-tu-telefono-Parte-2-2022.pdf

We’re celebrating the amazing women who are fighting to protect privacy and create a better world.

Get inspired ✨ #IWD2023

“Como é o caso do reconhecimento facial, que podem afectar em espcial pessoas pertencentes a grupos historicamente marginalizados, como mulheres”

We’re celebrating the amazing women who are fighting to protect privacy and create a better world.

Get inspired ✨ #IWD2023 #Shorts

“Privacy is important to me as it’s the right that upholds access to reproductive healthcare”

We’re celebrating the amazing women who are fighting to protect privacy and create a better world.

Get inspired ✨ #IWD2023

“Privacy is important to me because it allows me to live without being profiled, watched and influenced by outside actors”

We’re celebrating the amazing women who are fighting to protect privacy and create a better world.

Get inspired ✨ #IWD2023

“I first became interested in privacy when I learned my fellow latin American activists were being subjected to widespread state surveillance and living in fear”

We’re celebrating the amazing women who are fighting to protect privacy and create a better world.

Get inspired ✨ #IWD2023

Support Privacy International: https://pvcy.org/donateyt
Subscribe to our newsletter: https://pvcy.org/newsletter

You can find PI's guides to help you and your loved ones protect yourselves online: https://privacyinternational.org/act

If you're a streamer and you're interested in doing a similar stream with us please get in touch at socialmedia [at] privacyinternational.org

You can find guides on how to protect yourself and your love ones online here: https://privacyinternational.org/act

And keep an eye out for our podcast on Christmas Eve for more tips

The Data Interception Environment (DIE) is a tool that you can use to analyse how your data is being used by app developers and third parties. It allows you to see how apps are sending your data from your device back to the company or to third parties.

We’ve used to it research everything from low cost phones, to menstruation apps, and now we’re making it available so that you can do your own research on how your devices use your data.

Further Info

PI has made this tool available to assist individuals, researchers and organisations in developing a technical understanding of how their own data is being captured, processed and transferred by applications to 3rd parties. It should only be used to analyse data on a device, virtual copy of a device, or an app, which the user owns or is legally authorised to use.

The Data Interception Environment software is free to use in accordance with terms of the GNU General Public License. However, Privacy International’s brand, logo and name are copyrighted and protected from unauthorised use. PI cannot endorse any modifications that users or any other third parties make to PI’s Data Interception Environment and attribute to PI.

PI is a charity registered in England and Wales. We have developed our Data Interception Environment tool in accordance with the laws, regulations and standards that apply to us in England and Wales. Any individual, researcher or organisation who makes use of of PI’s Data Interception Environemnt is responsible for ensuring that they are acting lawfully and in accordance with the laws, regulations and standards which apply to them in the jurisdiction they are operating in or to which they are subject.

The Data Interception Environment (DIE) is a tool that you can use to analyse how your data is being used by app developers and third parties. It allows you to see how apps are sending your data from your device back to the company or to third parties.

We’ve used to it research everything from low cost phones, to menstruation apps, and now we’re making it available so that you can do your own research on how your devices use your data.

Further Info

PI has made this tool available to assist individuals, researchers and organisations in developing a technical understanding of how their own data is being captured, processed and transferred by applications to 3rd parties. It should only be used to analyse data on a device, virtual copy of a device, or an app, which the user owns or is legally authorised to use.

The Data Interception Environment software is free to use in accordance with terms of the GNU General Public License. However, Privacy International’s brand, logo and name are copyrighted and protected from unauthorised use. PI cannot endorse any modifications that users or any other third parties make to PI’s Data Interception Environment and attribute to PI.

PI is a charity registered in England and Wales. We have developed our Data Interception Environment tool in accordance with the laws, regulations and standards that apply to us in England and Wales. Any individual, researcher or organisation who makes use of of PI’s Data Interception Environemnt is responsible for ensuring that they are acting lawfully and in accordance with the laws, regulations and standards which apply to them in the jurisdiction they are operating in or to which they are subject.