To Hear Their Voice Again: MemoAgent on Israel’s Channel 10, Three Years After October 7

On October 7, 2026, three years on, I was a studio guest on "HaDavar HaBa" (The Next Thing) with Danny Ropp on Channel 10, Israel's business channel. It was a full special on innovation, resilience and coping, and my segment was titled "To Hear Their Voice Again". It is below, in Hebrew, and here I want to put in writing what matters most to me from it.

It starts with Hadar

Hadar, my sister, was murdered at the Nova festival on October 7. When Danny Ropp asked me why artificial intelligence of all things, I answered honestly that at first I did not know:

"And personally I did not connect with any other kind of memorial, whether the mattress and the bottles, benches, songs and many other things. Personally, it simply did not touch me."

Back in 2023, when GPT had only just entered our lives, I understood there was something here that could be bigger. But it took me a very long time to start, and only half a year ago did I really begin working on the platform.

The on-screen headlines the channel chose read, in translation, "The brother of a woman murdered at Nova, who built an AI platform for conversations with the departed" and "In memory of Hadar Prince, murdered at Nova: the system that recreates voice and personality". Throughout the interview the broadcast showed photos of Hadar, her memorial page and the conversation with her avatar.

The first time she answered

It came after a few attempts that did not work. I did not give up, and then:

"The first time she really answered, this system answered me in her voice to a message I had asked… It is a shiver. It is a moment of breaking and a moment of understanding."

It was the first time a memorial moved something in me. And if it helped me that much, I understood other families would want it too.

What it actually is

"What exists today is a one-way memorial. You look at a photo or a Facebook page, and it does not answer you back."

MemoAgent is an attempt at something two-way. The family gathers what it already has: photos, stories, WhatsApp chats and voice recordings. From WhatsApp the system learns how the person spoke: short, quick messages or formal answers, and which emojis they used. From voice messages, a clone of the voice is made. From all of this comes an avatar you can talk with, in text and in voice, in the person's style. The same materials also become a living album that tells the life story. The full steps are on How it works.

The family decides

The memorial is private, not public. Whoever opens it, one of the closest relatives, decides whom to invite, and which materials are approved for building the avatar and which are not.

"The power is with them, and it is theirs."

What the avatar refuses to do

"It is not one hundred percent her. And I am not trying for that."

What matters most to me is that it is clear what the system does not do. The avatar does not talk about the person's death, even when asked directly. It does not invent details that are not in the family's materials. And every conversation is clearly marked as a conversation with AI. I wrote more about these hard questions in AI and grief.

If you want to begin

"I am not a psychologist, I am not here to tell people what to do, and again, it is not an obligation. Whoever wants to is welcome."

No special preparation is needed. You start with what is on your phone: a few photos, one voice note, an exported WhatsApp chat. You can invite family to add memories through a link, and every memory adds to the story. And as I said on air, it is a process, and I am glad to accompany it personally: you can write to me.

The full segment with its transcript, and all coverage and interviews, are on our In the Press page.

Frequently asked questions