By Ariel Grossman, NoCamels -
Enraged by the denial of the Hamas massacre of hundreds of men, women and children in southern Israel on October 7, an Israeli programmer decided to create an AI chatbot that would educate people questioning the shocking truth of what happened.
Yuval Avidani, who has been working in information technology and cybersecurity for nearly two decades, developed the See The Truth chatbot after reading post after post on social media that denied the attacks, in which 1,200 people were killed and 240 more were abducted, ever took place.
A chatbot is an automated program that simulates a conversation between man and machine and is built as part of a website or an app. In this case, it is integrated into Avidani’s website devoted entirely to the war, with links to IDF data, information about the hostages and images that are so graphic that visitors wishing to view them are asked to confirm that they are over the age of 18.
Avidani believes that the power of generative AI is the best way to persuade skeptics of the facts, as well as show them the photographic evidence that is now difficult to access in other places online after the images were deleted from Telegram and the platform restricted Hamas channels.
“I wanted to do something to contribute to my country and to my people by explaining – and showing – the horrors that took place,” Avidani tells NoCamels.
“The idea is to shock them and convince them, because I hear more and more people saying that it isn’t true,” he says.
“When you ask questions and you get these images, you can’t say this didn’t happen.”
To create the chatbot, Avidani first trawled through the Telegram channels affiliated with Hamas and the Hezbollah terror group in Lebanon.
From these channels, he saved brutal images and videos of Israelis who had been murdered and their bodies desecrated.
Now with all the images uploaded, Avidani spends several hours each day updating the bot with new and credible information about the events of the ongoing war that Israel declared on Hamas in Gaza in the immediate aftermath of the attack.
To reach as many people as possible around the world, Avidani says that he used the AI to allow the chatbot to answer queries in any and all languages.
Generative AI bots such as ChatGPT normally have restrictions and filters to prevent users from receiving answers to violent questions, such as methods of killing people or creating explosives at home. To ensure that those seeking answers about the events on and following October 7 do not receive restricted information, Avidani overrode the filter limitation by writing specific code into his Microsoft-powered chatbot.
Furthermore, the AI would at first automatically pull in information using keywords, but this too was problematic as it could not distinguish between impartial news items and subjective editorials and opinion pieces, leading to it providing biased information. This too was overcome by Avidani’s own specialized coding.
Now, the chatbot not only provides links to trustworthy websites about the horrors of the attack, but also shows images and videos that back up those reports.
And just as there have been multiple cases of people around the world tearing down posters of the Israelis being held hostage by Hamas, the bot too was immediately targeted when it first began operating about a week after the war began.
“All of a sudden, I received an alert that someone was trying to send 40,000 messages per second,” says Avidani. “This means that someone was running an automated tool and attacking my website to abuse and break it.”
After rewriting the code again and adding various security measures, Avidani was able to get the bot up and running again.
Since then, thousands around the world have used it, in particular from the United States and Israel, as well as hundreds from Canada, Australia, the United Kingdom and Germany.
The bot has been shared by Avidani and others on social media and messaging apps, and he has also promoted it through his own Hebrew-language website for AI enthusiasts.
Avidani says that some of the visitors have used the bot to receive accurate and concise AI-generated responses, so they can confidently respond to anti-Israel posts and comments circulating online.
The overwhelming majority of visitors, he says, are simply curious to understand the details and full events of October 7: how Israeli intelligence failed; how Hamas terrorists infiltrated the southern border communities; and to see the proof of their rampage.
To provide more evidence, Avidani next plans to upload audio testimonies of Israelis who were living in the Gaza border community, which will also be translated into dozens of languages.
In fact, Avidani began working on this feature when a relative of one of the hostages approached him.
“They told me that they wanted to tell the world what happened, and that they needed their stories to be translated into as many languages as possible,” he says.
Avidani has also created a version of a chatbot for the Telegram app, aiming to make the information accessible to people using different forms of social media.
“I don’t enjoy showing the world the people of my country being slaughtered,” he says.
“My only goal is to make people stop believing the lies of Hamas and the fake news they are spreading all over, and showing what they actually did.”
It is for this reason, says Avidani, that the chatbot is dubbed “See the Truth.”
“I don’t want you to just understand the truth,” he says. “I want you to see it. And It’s horrific.”