Pi (Inflection)
- Your data trains their AI unless you switch it off in settings.
- The policy rules out using your data for targeted ads.
- They say they do not sell or share your data.
- Staff may review your content; the policy does not describe an opt-out.
Chats can train models unless you turn that off; they keep a broad license, can close you anytime, and cap payouts at $100.
Agent Trust Hub uses AI and may make mistakes. Review reports and confirm their contents before relying on them.
Collected
Account info, chats and voice, device IDs, cookies, and sometimes location, contacts, and partner data.
Read the exact line
Privacy policy
“Depending on which Services you use, the types of information that we collect directly from you include your name, email address, “inputs” consisting of text, voice, audio, or other materials”
Training
They run the app, personalize it, and train AI models unless you turn training off in Account settings.
Read the exact line
Privacy policy
“Provide, maintain, and improve the Services, including to develop and train our AI models / large language models that power our Services. You may opt out of model training in your Account settings page;”
Sharing
Vendors (hosting, moderation, marketing) and, if required, law enforcement; they say they do not sell your data.
Read the exact line
Privacy policy
“We do not engage in “sales” or “sharing” of personal information or process personal information for “targeted advertising” purposes as those terms are defined by U.S. state privacy laws.”
Kept
Account data while you stay; deleted chats usually 15 days; they keep outputs indefinitely for limited uses.
Read the exact line
Privacy policy
“For users of Pi, we generally retain inputs for at most 15 days after user deletion. In some cases, we retain inputs for longer, such as if needed to address fraud and abuse”
Controls
You can delete the account in the app or type!delete, and turn off model training in Account settings.
Read the exact line
Privacy policy
“Users of Pi may delete their information at any time within the Service by either (1) navigating to their profile page within the Service, selecting Account, and clicking on the “Delete my account” button; or (2) typing “!delete” within the chat.”
Fine print
They can close you anytime.
Read the exact line
Terms of service
“Inflection AI is free to terminate or suspend access to your use of our Services at any time and for any reason at our discretion.”
Expand “Read the exact line” to see the source alongside the explanation.
What you can turn off
The controls and opt-outs their own documents describe, and where they say to find them.
Turn off model training
Account settings page
Turn off voice emotion inference
Account settings page
Delete your account
Profile page → Account → “Delete my account”, or type!delete in chat
If a switch is not where they say, the deletion request above still applies.
Line by line
The lines that matter most, worst first.
Your chats are used to train their AI unless you switch that off in settings.
Read the exact line
Privacy policy
“Provide, maintain, and improve the Services, including to develop and train our AI models / large language models that power our Services. You may opt out of model training in your Account settings page;”
For example, a private vent about work could help teach the model unless you turn training off first.
If someone sues them over how you used Pi, you may have to pay their legal bills.
Read the exact line
Terms of service
“you will and Inflection AI and the other Inflection AI Parties from and against any and all claims, liabilities, damages (actual and consequential), losses and expenses (including attorneys’ fees)”
For example, if a third party claims your chat content caused harm, you could be on the hook for Inflection’s lawyer fees.
If they mess up, they say they will not pay you more than $100 in most cases.
Read the exact line
Terms of service
“FOR ANY TOTAL, AGGREGATE DIRECT DAMAGES, COSTS, LOSSES OR LIABILITIES IN EXCESS OF ONE HUNDRED ($100) U.S. DOLLARS.”
For example, if a bug leaked your chats, they claim they would owe at most one hundred dollars.
They can shut off your account without promising a warning.
Read the exact line
Terms of service
“Inflection AI is free to terminate or suspend access to your use of our Services at any time and for any reason at our discretion.”
For example, you could lose access to your chat history overnight if they decide to close the account.
Pi’s replies are not fully erased on a short timer; they can keep them forever for the uses in the terms.
Read the exact line
Privacy policy
“We retain outputs indefinitely for the limited purposes described in Section 4 of our Terms of Service.”
For example, even after you delete chats, copies of Pi’s answers may still sit in their systems.
They can guess how you feel from what you say, including voice, unless you turn the voice-emotion setting off.
Read the exact line
Privacy policy
“We may use your inputs to make inferences about your emotional state, tone, sentiments, or other characteristics. You may opt out of using voice to infer emotions in your Account settings page.”
For example, a shaky voice note could be treated as a signal that you are upset.
If you log in with Google, Apple, or similar and allow it, they can pull contacts, mail, calendar, or files.
Read the exact line
Privacy policy
“If you provide us with the relevant permission, we may have access to information in your contacts, emails, calendar, or documents on such third-party platform.”
For example, connecting Google could let them see names from your address book.
They say they do not sell your info or use it for targeted ads under those U.S. state definitions.
Read the exact line
Privacy policy
“We do not engage in “sales” or “sharing” of personal information or process personal information for “targeted advertising” purposes as those terms are defined by U.S. state privacy laws.”
For example, they claim they will not sell a list of Pi users to an ad network under those laws.
You can wipe the account yourself in the app or by sending!delete in chat.
Read the exact line
Privacy policy
“Users of Pi may delete their information at any time within the Service by either (1) navigating to their profile page within the Service, selecting Account, and clicking on the “Delete my account” button; or (2) typing “!delete” within the chat.”
For example, typing!delete in Pi should start deleting your account without emailing support.
Words to know
Legal words from the lines above, in plain English.
- inputs and outputs
- Inputs are what you type, say or upload to an AI; outputs are what it gives back to you. For example, the question you ask a chatbot is an input and its answer is an output, and both may be stored.
- opt out
- Something is on by default and stays on until you find the setting and turn it off. For example, your chats may be used for training unless you go into settings and switch it off.
- model training
- Using your content and conversations as examples to teach an AI system, which can then echo them in future answers. For example, a story you write in a chatbot may be studied by the company to make the next version of the bot.
- personal data
- Any information that is about you or can be linked to you, from your name to your phone's ID. For example, your email, your IP address and the list of apps on your phone are all personal data.
- targeted advertising
- Ads chosen for you based on what you have done across other apps and websites, not just this one. For example, you look at shoes in one app and see shoe ads in a totally different app the next day.
- retention
- How long a company keeps your data before deleting it. For example, a policy might keep your messages for 30 days after you delete them, or for as long as it likes.
- indemnify
- You promise to pay the company's legal costs and losses if your use of the service gets it sued. For example, if you post a song you do not own and the label sues the app, you owe the app's lawyer bills.
- profiling
- Building a picture of who you are, what you like and what you might do next from your data. For example, an app may guess your income, health or politics from what you watch and tap.
- third parties
- Any company or person other than you and the app, such as advertisers, partners or analytics firms. For example, an analytics company that receives a record of every screen you tap is a third party.
How we got here · grade C · score 42/100 · 13 of 15 policy answers backed by a verified quote · 2 not stated
What does the app collect beyond what it needs to work?
Besides account details and chats, they collect device IDs, IP, cookies, optional precise location, and, with permission, contacts, emails, calendar, or documents, plus public web info and emotion inferences.
Does it record your voice, face or body, and what happens to that?
Voice and audio you send are collected as inputs, can be used to guess emotions unless you opt out, and inputs are used to train models unless you turn training off.
Are your chats and uploads used to train AI models, and is that off by default?
They use your information to train the models that power Pi, and they say you can turn that off on the Account settings page.
Can employees or contractors read your conversations, and when?
They say they use automated tools plus human review for safety, and they share data with content-moderation vendors. The policy does not describe an opt-out from that review.
Are you profiled or tracked for advertising?
They say they do not process personal information for targeted advertising as U.S. state privacy laws define it. Cookies and marketing vendors are described, but that alone is not treated as targeted ads here.
Do they sell or share your data, and can you opt out?
They explicitly say they do not sell or share personal information as those terms are defined by U.S. state privacy laws. They still give data to vendors who work for them.
What rights do they take over what you type and what the AI makes?
You own inputs and outputs, but you give them a perpetual, irrevocable, sublicensable worldwide license to operate the service and to train and improve models. That license survives when you leave.
How long do they keep your data after you delete it, and can you delete it in the app?
You can delete in the app or with!delete. Inputs are generally gone within 15 days after deletion, but they keep outputs indefinitely for limited purposes and may keep inputs longer for fraud, security, or law.
Does it build a lasting memory or profile of you, and can you see, edit or turn it off?
Not stated in the documents.
Do they commit to basic security, and have they leaked data?
Not stated in the documents.
Is there a real age gate, and are teens protected?
They say the services are not for anyone under 18, but they only ask you not to use them if you are under 18. No extra age check is described.
Will they tell you when the rules change, and is the policy specific?
The privacy policy is dated. They update the date for changes and may give extra notice for material privacy changes. Terms say continued use after they post a new date means you accept.
Can they close your account without warning, and can you get your data out first?
They can suspend or end your access at any time for any reason at their discretion. You can close the account yourself, but they do not promise notice before they cut you off.
If something goes wrong, who pays?
The service is “as is,” they cap most payouts at $100, and you must cover their legal costs if a third party sues over your use or your content.
Do subscriptions renew on their own, and can you get a refund?
These documents do not describe a paid subscription, auto-renewal, or refunds.
The privacy policy is 11,991 words at a college graduate level (Flesch reading ease 38.8); the terms are 5,901 words at a professional or legal expert level.
