AI Wrote the Obituary — Is That Okay? A Gentle Guide to AI, Grief, and Etiquette

Everyday life Story8 min read·Updated July 11, 2026
The short answer

Funeral homes increasingly use AI drafting tools without telling families, which can introduce factual errors or a tone that doesn't match your loved one. You have the right to review, edit, or write the obituary yourself. For condolence messages, AI can help you find words when grief leaves you stuck — but the message lands better when you rewrite it in your own voice and add one specific memory.

Death comes without warning, and it often leaves the people it touches without words. A grief-struck family member is asked to produce a complete, accurate life history of their loved one within hours of their passing — and increasingly, the funeral home quietly hands that task to an AI tool, sometimes without mentioning it at all. At the same time, more people are turning to AI when they themselves can't find the words to reach out to someone who is grieving. Neither of these is wrong, exactly — but both carry risks worth knowing.

Part One: When the Funeral Home Uses AI

A Washington Post investigation published in August 2025 found that funeral homes across the United States were using AI tools — including software called Passare — to draft obituaries at scale, often pulling from basic intake forms filled out while families were in shock. Families who later noticed errors reported finding wrong birth years, invented hobbies, and phrasing that bore no resemblance to who their loved one actually was. Most of those families said the funeral home had never mentioned that AI was involved.

This doesn't mean AI-assisted obituary drafting is inherently bad — a compassionate first draft, produced quickly, can genuinely help a grieving family who doesn't know where to start. The problem is when the draft goes out without the family seeing it, or when errors make it into print or a permanent online memorial.

You have more control here than you may realize.

What to Ask the Funeral Home

When you first sit down — in person or by phone — with a funeral home to arrange a service, it is completely reasonable to ask: "Will you be writing the obituary, or will we?" and "If your team drafts it, can we review it before it's published?"

A funeral home that uses AI tools should be willing to show you the draft. If they're not, or if they've already published something you didn't approve, that's worth addressing directly. Most funeral homes depend on reputation and word of mouth — they genuinely want families to be satisfied, and will correct errors if you ask.

How to Review an AI-Drafted Obituary

When the draft comes to you, go through it in two passes.

First pass — facts only. Check each of the following:

  • Full name, including middle name or maiden name
  • Birth date and place
  • Death date
  • Names of surviving family members (spouse, children, siblings, parents)
  • Names of family members who died before them (predeceased)
  • Where they lived, and for how long
  • What they did for work, and when they retired

These are the details most likely to contain errors when an AI tool works from a brief intake form. A wrong birth year, a misspelled family name, or a missing sibling will matter enormously to the people who read this in the years ahead.

Second pass — tone. Read the whole thing aloud. Does it sound like someone who knew this person? AI-drafted text tends to fall back on generic formulations — "devoted parent," "beloved by all," "will be deeply missed" — which are true of almost everyone and feel true of no one in particular. You have every right to change the language so it sounds like your family actually talking about this specific person.

Prompt you can use: "Help me revise this obituary draft so it sounds more like Name. Here are three things that were most true about them: your list. Can you suggest sentences that bring those in?"

Part Two: When You Can't Find the Words to Reach Out

Sending a condolence message is one of the hardest kinds of writing there is. You want to say something true, something that acknowledges real loss — and grief has a way of jamming the words before they make it to the page.

AI can help. Asking ChatGPT or Gemini to help you draft a condolence note is not cold or lazy — it's a practical way to get unstuck when you've been staring at a blank screen for twenty minutes. But the way most people do it (send the AI draft more or less as-is) often doesn't land the way they hope.

Why AI Condolence Messages Often Feel Generic

Etiquette writers have long noted that what makes a condolence message feel real is one specific, personal detail — something that shows you actually knew the person who died, or truly know the person who is grieving. "I still think about the way he laughed at his own jokes before he finished telling them" is more comforting than "Your loss is immeasurable." AI tends to produce the second kind of sentence.

The people receiving condolence messages are especially attuned to this. Grief makes everything feel heightened — both the kindness that lands and the emptiness that doesn't. A message that reads like a template, even a technically warm one, can feel like no one really stopped to think about the person who died.

The Middle Path That Works

Use AI for the first draft. Don't send the first draft.

After AI gives you something to work with, do this:

  1. Read it and mark anything that's actually true about your relationship to the person or their loss.
  2. Cut the generic phrases ("words cannot express," "in my thoughts and prayers" unless you genuinely mean it, "they are in a better place" unless you know the family's beliefs).
  3. Add one sentence — just one — about a specific memory or quality. It doesn't have to be poetic.
  4. Read it back to yourself. If it sounds like you, send it. If it still sounds like a template, try once more.

Prompt to try: "Help me write a condolence note for Name, who lost her father last week. I knew him briefly — he coached my son's soccer team for two years and always stayed after practice to help kids who were struggling. I want to tell her I'm thinking of her without it sounding like a greeting card."

That last line matters: telling AI what you want to avoid is often more useful than describing what you want.

When It's Better to Call

Some losses are close enough that a written message isn't quite right — a childhood friend, a sibling's spouse, someone you've known for decades. In those cases, don't let the difficulty of finding words become a reason to stay silent. A short, imperfect voicemail or a handwritten note that arrives a few days late is better than a polished AI message that arrived immediately but felt like nothing. If you're unsure whether your message hits the right tone before you send it, the tip on checking your message's tone before you send it can help — but that works best when the words in the message are already yours.

What to Watch Out For

Factual permanence. Obituaries often become part of the permanent record — newspapers archive them, genealogical databases index them, families save them. An error that seems minor ("born in Toledo" instead of "born in Dayton") can confuse family records for generations. Review carefully before anything is published.

AI filling in what it doesn't know. Some AI drafting tools, when given sparse information, will plausibly complete details — a vague mention of "community involvement" might become "served on the school board" because that seemed likely. That's an invention, not a mistake, and it's the hardest kind to catch. Read the draft against what you actually know about the person, not just against what sounds plausible.

Religious and cultural language. AI defaults to familiar phrases that may not match the family's beliefs or traditions. An obituary for a secular family that says "she has gone to her eternal rest," or a condolence note that promises "heaven's welcome" to a family of a different faith, can land poorly. You're allowed to take this language out.

Privacy. Publishing an obituary puts personal information — names, relationships, locations — into public view. Some families prefer to omit surviving children's ages, home cities, or other details that might be useful to people who weren't meant to be part of the service. That's a reasonable call, and you don't need the funeral home's permission to ask for less detail.

What to Try Next

If writing about a life — whether an obituary, a eulogy, or something longer — has made you want to preserve more of your loved one's story, Let AI Help You Write a Life Story or Family Memoir walks through how to approach that with care. And if you're looking for help with a different kind of sensitive message — an apology, a thank-you, something you've been putting off — AI Message Templates: How to Start Messages You've Been Avoiding covers the same principle of using AI to start and then making it yours.

Published July 11, 2026 · Updated July 11, 2026How we test →

Frequently asked questions

Can I ask the funeral home to let me write the obituary myself?
Yes, and most funeral homes will welcome it. The obituary belongs to your family — the funeral home's job is to help you publish it, not to decide what it says. If they offer to draft it for you (with or without AI), you can simply say you'd like to write it yourself or have a family member do it. Many families find writing the obituary a meaningful part of saying goodbye, not a burden.
What should I look for when I review an AI-drafted obituary?
Start with the facts: full name (including middle name or maiden name), birth date, death date, birthplace, names of surviving and predeceased family members, where the person lived, and what they did for work or in the community. These are the details AI tools are most likely to get wrong, especially if the funeral home pulled information from a brief intake form. Then read for tone — does it sound like the person? A draft that calls a lifelong union worker 'a beloved entrepreneur' needs to be corrected before publication.
Is it wrong to use AI to write a condolence message?
It's not wrong — it's a tool, and people use tools when they're struggling. Grief can leave you genuinely speechless, and asking AI to help you find the right words is not the same as not caring. The issue isn't using AI for a first draft; it's sending that draft untouched. A message that reads like a template — even a warm one — can feel impersonal to someone who is grieving. Take what AI gives you, then rewrite it in your own words and add one thing you actually remember about the person.
Will people be able to tell if my condolence message was written by AI?
Often, yes — not because they run a detection tool, but because AI condolence messages tend to use similar phrases ('I hold you in my thoughts,' 'words cannot express') and lack the specific, personal details that make a message feel real. Grief is one of the situations where generic language is most noticeable. The antidote isn't to avoid AI; it's to use AI as a starting point and then make it yours.
What if the funeral home already published an obituary with a factual error?
Contact the funeral home directly and ask for a correction. Most obituaries published online can be updated, and reputable funeral homes will do this at no extra charge when there's a clear factual error. If the obituary has already run in print, many newspapers publish a correction or a follow-up notice. Document the error first — take a screenshot — before you call, so you have a clear reference when asking for the change.
Do I have to tell people if I used AI to help with a condolence message?
No — you're not obligated to disclose how you drafted a personal message. What matters is that the message reflects your genuine care and contains something true about your relationship with the person. If you used AI to find your words and then made them your own, that's no different from someone who looks up a poem to share, or asks a friend what to say. The sincerity is in the sending.
Radim S.
Founder & editor

Radim is a software developer who spends his days building with AI and his evenings explaining it to family members who don’t care how it works — only what it can do for them. Every guide is tested by hand before it’s published.