72 uses automated systems to provide requested features and safety controls. Automated output can be incomplete or wrong and must not be treated as professional advice or a guarantee about another person.
01Current uses
AI or automated models may assist liveness/face consistency, photo and media moderation, voice transcription, text translation, profile-about drafting, safety signals, and abuse detection.
02Human control and review
Users decide what to send and can report errors or harmful outcomes. Safety enforcement may include human review; rights to contest or request review apply where required.
- Outgoing translation and voice transcription can use a token allowance. Counts come from confirmed provider usage, including instructions, input and output. Example message counts are not guarantees. Original-only messages are not processed for translation or transcription and are not backfilled after a later purchase.
03No model training
User data is not used to train a general-purpose AI model for 72 or a third party. A future training purpose requires a new notice and separate lawful basis or consent before activation.
04Providers and minimization
Only the data needed for the requested operation should be sent to the listed AI/cloud processor. Temporary OpenAI API files are deleted after processing; provider service logs remain governed by the applicable processor terms and retention controls.
05High-level principle
AI is assistive; final decisions remain with users. AI outputs may be inaccurate, incomplete, biased, delayed, or contextually wrong. 72 does not guarantee any match outcome, legal identity, official KYC, gender, age, location, intention, compatibility, relationship outcome, or real-world safety.
06Verification, liveness, and profile consistency
AI or automated systems may help assess camera frames, live verification signals, declared-information consistency, real-camera/replay risk, and profile photo consistency. These are not official KYC, legal identity verification, guaranteed gender verification, guaranteed age verification, guaranteed location verification, or a safety guarantee. Some face/camera data may be sensitive, special-category, or biometric data in some jurisdictions; the Privacy Policy and applicable law apply.
07Profile photo and media moderation
Profile photo and profile media may be reviewed by automated moderation tools and/or human/admin review. Content may be allowed, hidden, disabled, quarantined, removed, or reviewed. Moderation can be wrong; support or appeal may be available where required or appropriate.
08Profile text assistance
If enabled, AI may help draft or improve user profile/about text. The user remains responsible for what appears on the profile. AI-generated profile text must not be used to deceive, impersonate, or misrepresent identity, intent, age, gender, location, or relationship status. This is not professional advice.
09Human/admin review
Human or admin review may support safety, reports, moderation, support, account integrity, and appeal processes. Details may be limited for anti-abuse, security, privacy, or legal reasons. Serious restrictions may be eligible for support or appeal where required or appropriate.
10Service providers
72 may use Firebase/Google Cloud infrastructure, OpenAI or other AI providers, speech/transcription/translation providers, moderation/safety providers, analytics/crash/log providers where applicable, and Apple/Google platform services where applicable. International transfers are handled under the Privacy Policy.
11Limitations
72 does not provide:
- Legal identity verification or official KYC.
- Guaranteed gender, age, or location verification.
- Guaranteed compatibility, safety, or relationship outcome.
- Guaranteed translation or moderation accuracy.
- Emergency service or professional advice.
- A guarantee that media cannot be externally recorded.
12Translation, speech and video subtitles
Text translation can process a message and language context, including sender and recipient grammatical-gender context where relevant. Voice features can process audio to create a transcript and translation. Video subtitle features can process speech, timestamps and text to generate editable captions. Results can miss speech, split timing poorly or misunderstand names, slang, negation, tone or intent. Review important output, compare the original where available and correct generated captions before using or sharing them. AI output is not a verified statement by the speaker or professional advice.
13Feature-specific inputs and provider access
Depending on the feature you request, the required text, audio, image, verification frames or associated language context may be processed by an AI provider. Verification, photo consistency and moderation have different purposes from chat translation. We limit processing to the relevant workflow and do not describe all-chat surveillance, compatibility scoring, official identity checks or relationship guarantees as features of 72. The Privacy Notice explains providers, consent choices, retention and international processing.