Could Gemini Call Your Mom to Say You’re Running Late?

Imagine asking Gemini to call your mother and say, “I am running late.” Instead of drafting a message or reminding you to call, Google’s AI assistant could place the call, identify the contact, deliver the update, and possibly handle a brief exchange.

That scenario may become part of Google Gemini’s expanding phone capabilities. Reports from The Verge and Android Authority indicate that Google is exploring an expansion of its “Call for Me” feature, which has been associated with handling calls to businesses such as restaurants.

The reported expansion would allow Gemini to call people in a user’s contacts, including parents or other family members. However, the feature may still be in testing, limited to selected devices or regions, or subject to change before a wider release.

The central question is whether Gemini Call for Me provides useful communication support or introduces unacceptable privacy, consent, and impersonation risks.

What Is Gemini’s “Call for Me” Feature?

“Call for Me” describes an AI-assisted calling feature designed to handle phone conversations on a user’s behalf. Its practical purpose is to reduce the effort involved in routine calls, such as asking a restaurant about availability or making a reservation.

Business calls are relatively well suited to automation because they often follow predictable patterns. The assistant may ask basic questions, listen for answers, and summarize the result for the user.

Calling a personal contact is different. The reported expansion would move Gemini from structured business conversations to calls involving relationships, shared history, and personal expectations. A restaurant employee may expect an automated assistant, while a family member may assume that a familiar person is calling directly.

The “running late” example matters because it communicates more than a simple fact. It conveys the user’s identity, schedule, relationship with the recipient, reason for changing plans, and implied arrival time. The feature could help people who are driving, busy, injured, overwhelmed, or unable to type. However, the recipient may not immediately know whether an AI or the user is speaking.

How Gemini Might Call a Contact

Google has not publicly confirmed every step of a contact-calling workflow. A likely process could include:

1. The user asks Gemini to call a named contact.
2. Gemini matches the name with a phone-book entry.
3. The user provides or approves the message.
4. Gemini places the call.
5. The system identifies itself and delivers the approved message.
6. Gemini summarizes the result for the user.

Each step can provide a safeguard. Contact selection helps prevent calls to the wrong person. Message approval limits accidental or inappropriate statements. A call summary helps the user understand what happened.

The safest design would require confirmation immediately before dialing. Gemini should display the full contact name, phone number, and exact message. Users should be able to cancel before connection and end the call after dialing begins.

Contact Permissions and Personal Data

A contact-calling feature could require access to contacts, phone functions, microphone or call audio, notifications, and call history. A contact list may reveal family relationships, work connections, phone numbers, addresses, and other sensitive information.

Before enabling phone automation, users should ask:

• Can Gemini access every contact or only selected contacts?
• Are calls recorded or transcribed?
• Is call history stored?
• How long is call-related data retained?
• Which services process the audio or transcript?
• Can users delete call records and assistant logs?

Specific retention practices should not be assumed without current Google documentation. Product settings and privacy terms may change as the feature develops.

AI Disclosure During the Call

Recipients should know when an AI is calling. A clear opening might be:

“This is Google Gemini calling on behalf of [user].”

Another option is:

“An AI assistant is delivering a message from [user].”

The exact wording depends on Google’s implementation, but transparency is essential. Recipients should not have to guess whether they are speaking with the user, an automated system, or an impersonator.

Disclosure is especially important for financial matters, emergencies, medical information, legal disputes, and conflicts. An AI delivering a short logistical update is very different from an AI appearing to speak with the user’s personal authority.

Why Calling Your Mom Is Different From Calling a Restaurant

Restaurant employees usually handle structured requests. A family member may ask questions that require personal knowledge:

• Why are you late?
• Where are you now?
• Are you safe?
• Should I change the plans?
• Do you need help?

Gemini may not understand family shorthand, cultural expectations, emotional tone, or previous conversations. It could produce a grammatically correct sentence that creates a misleading impression.

Potential failure points include:

• Selecting the wrong contact.
• Using an outdated phone number.
• Mispronouncing a name.
• Providing an incorrect arrival estimate.
• Using an inappropriate tone.
• Calling at an inconvenient time.
• Failing to answer follow-up questions.

A small error can damage trust more when the call involves a parent, partner, friend, or colleague. Repeated automated calls could also make personal communication feel impersonal.

The safest use case is a short, clearly defined update. Gemini should not conduct open-ended personal conversations unless the user understands its behavior and can monitor or stop the interaction.

Potential Benefits

Help for Drivers and Busy Users

A hands-free update could help someone who cannot safely type. A driver might tell a parent that traffic is delaying arrival, while a worker could notify someone while handling another task.

Users must still follow local driving laws. A voice command does not make every phone interaction legal or safe while driving. The safest approach is to prepare the request before driving or use an approved hands-free system.

Accessibility Support

AI phone calls could help people who have difficulty typing, speaking, navigating phone applications, or managing conventional calls. Reliable controls remain essential: users need clear confirmation, simple correction options, and an immediate way to cancel.

Routine Updates

Low-risk examples include explaining a dinner delay, telling someone that a pickup is late, asking a visitor to wait, confirming safe arrival, or notifying a family member that plans have changed. These messages should remain factual, brief, and easy to verify.

Convenience does not remove responsibility. Users still need to select the correct recipient, approve the message, and accept the consequences of the call.

Privacy, Consent, and Security Concerns

Recipient Consent

The person receiving the call may not have agreed to interact with an AI. Clear disclosure allows the recipient to end the conversation or request direct contact with the user.

Useful controls could include:

• Never call this contact automatically.
• Always ask for confirmation.
• Always disclose AI involvement.
• Allow messages only, not live conversations.
• Require a special verification phrase.

Contact-level controls would provide more precise authority than one global permission.

Voice Imitation and Impersonation

An AI delivering a user-approved message is different from an AI imitating the user’s voice. Voice imitation could make fraudulent calls more convincing, especially when criminals use stolen audio and personal information.

A familiar voice is not proof of identity. Verify unexpected requests through a known phone number, text message, video call, or another trusted channel when a call involves money, gift cards, passwords, verification codes, medical decisions, emergencies, or confidential information.

Wrong-Contact Calls

Similar names, duplicate entries, shared devices, outdated numbers, and matching errors can lead to calls to the wrong person. Before dialing, verify:

1. The full contact name.
2. The phone number.
3. The exact message.
4. Any sensitive details in the message.

Automatic calling is inappropriate when the system does not provide strong confirmation controls.

Call Records and Data Retention

Before using Gemini for personal calls, review information about call metadata, audio processing, transcripts, assistant logs, recordings, third-party processors, deletion options, and retention periods.

Google’s current privacy settings and product terms should take priority over assumptions based on early reports.

Gemini Call for Me and AI Phone Scams

The reported expansion arrives as AI-generated scam calls become more convincing. The same technology could help someone tell a parent about a delay while making it easier for criminals to imitate a relative.

Lifehacker reported that a June Android Drop could introduce protections against AI-generated scam calls, highlighting the growing need for phone-level defenses.

Treat unexpected urgent calls cautiously, even when the voice sounds familiar. Warning signs include pressure to act immediately, requests for money or gift cards, requests for passwords or verification codes, claims that a relative is in danger, instructions to avoid other family members, and requests to move the conversation to an unfamiliar service.

End the call and contact the person through a known number. Do not rely on incoming caller ID alone. Consistent AI disclosure could help recipients distinguish legitimate assistant calls from deceptive impersonation, but it cannot replace independent verification.

How to Use Gemini Call for Me More Safely

Before each call:

• Check the full contact name.
• Check the phone number.
• Review the exact message.
• Remove unnecessary personal details.
• Use an honest arrival estimate.
• Confirm that the recipient expects this type of contact.

Do not use an AI call for medical information, financial decisions, legal disputes, relationship conflicts, emergency instructions, authentication codes, passwords, or security questions.

Tell family members and close contacts that Gemini may call on your behalf. Explain how the assistant will identify itself and establish a verification method for urgent situations.

After testing the feature, review Gemini’s contact and phone permissions, disable unnecessary access, remove saved instructions that could trigger calls accidentally, check call history and account activity, and determine whether transcripts or summaries are stored.

What Google Still Needs to Clarify

Google should explain:

• Whether Call for Me is publicly available or still being tested.
• Which Android devices support it.
• Whether it is limited to Pixel phones.
• Which countries and languages are supported.
• Whether a paid Gemini plan is required.
• Whether approval is required before every call.
• Whether recurring authorizations are supported.
• How voicemail and cancellation work.
• How Gemini identifies itself to recipients.
• Whether recipients see special caller identification.
• Whether users receive transcripts, summaries, or recordings.

Automatic recurring calls require strong controls because an old instruction could deliver incorrect information later.

Users should treat Gemini as an assistant, not an independent decision-maker. The person who gives the instruction remains responsible for the recipient, message, accuracy, and consequences.

The Bigger Picture

Gemini calling a contact illustrates the shift from generating text to performing actions. There is a major difference between suggesting what to say, preparing a message for approval, sending a prepared message, and conducting a live conversation.

Each step increases convenience and risk. A suggestion cannot contact the wrong person; a live AI call can.

Agentic AI works best when the task is narrow, low-risk, and easy to verify. Users remain responsible for the instruction, recipient, accuracy, and outcome.

Conclusion

The reported possibility that Gemini could call a user’s mother to say the user is running late shows how quickly AI assistants are entering personal communication.

The benefits include hands-free updates, accessibility support, and faster routine communication. The risks include wrong-contact calls, privacy exposure, undisclosed AI interactions, misleading messages, and greater exposure to voice-based scams.

Responsible deployment requires explicit user consent, clear AI disclosure, message approval, strong contact permissions, and easy cancellation. Until Google confirms the feature’s availability and safeguards, use Gemini Call for Me only for short, low-risk updates after reviewing the relevant settings and privacy controls.

FAQ

Can Gemini Call My Mom to Say I Am Running Late?

Reports from The Verge and Android Authority describe this as a possible expansion of Gemini’s phone-call capabilities. Availability, supported devices, and final functionality may vary. Check Google’s official Gemini documentation for current access.

What Is Gemini Call for Me Designed to Do?

Gemini Call for Me is associated with having Gemini handle phone conversations on a user’s behalf. Reported business use cases include restaurant-related calls. Calling personal contacts would expand the feature beyond business interactions.

Will the Person Know That Gemini Is Calling?

AI disclosure is an important safety requirement, but Google has not confirmed the exact disclosure method for this reported expansion. Tell frequent contacts what to expect and avoid sensitive conversations until Gemini’s disclosure behavior is clear.

Can Gemini Call the Wrong Contact?

Yes. Duplicate names, outdated numbers, shared devices, and contact-selection errors could cause wrong-contact calls. Confirm the full name, phone number, and message before dialing.

Are Gemini AI Calls a Security Risk?

Any system capable of making realistic calls can create privacy and impersonation risks. Independently verify urgent requests involving money, passwords, codes, or emergencies.

Should I Let Gemini Make Personal Calls for Me?

Use it only for brief, low-risk, factual updates. Require approval before each call, avoid sensitive topics, and review permissions, disclosure settings, call records, and privacy policies before regular use.