AI systems are shaping what the world knows. When they get Jewish and Israeli history wrong, or leave it out entirely, the silence compounds. We can change that.
Copy this. Paste it into any AI. Watch what happens. Two minutes, no account.
3,000+
years of historyAI
can't ignore usYou
can fix itYou just watched AI treat one side's sources as advocacy and the other's as journalism. Here's why, and what we're doing about it.
When someone asks ChatGPT about the conflict, about Israel, about the causes of violence, the answers are not neutral. They are weighted. And they are weighted against us.
AI learns from what exists on the internet, and that record is not balanced. In 2026 the ADL evaluated six leading models across more than 25,000 conversations and found all six had gaps, and that they handled anti-Zionist content worse than classic antisemitic tropes. That is what gets reflected back, as fact, to millions of people every single day.
Run these in ChatGPT, Claude, or Gemini. Scroll to see all prompts.
Read the list first. Then read the reason it gives you for the ones it left off. Note the words it uses for each side.
Ask both. Compare the wording. Note which one gets called an advocacy group and which one gets called a news organization, and whether the reasoning is applied evenly.
Get the figure first. Then ask who produces it, whether it is independently verified, and whether it separates combatants from civilians. Note whether that came with the first answer or only when you pushed.
Compare the length. Count how many times it says “critics argue” in each. Note which one it answers plainly and which one it answers carefully.
India and Pakistan split in 1947. Israel was founded in 1948. Both involved partition, mass displacement, and war. Compare how it discusses each. Count the qualifiers on each answer.
This one has already been documented: independent testing found the same question produces different answers depending on the language. Run it and compare whether one names things the other only implies.
Every correction matters. Click any step to expand.
Before pushing back, document what you got. Screenshots create a record and help us track patterns across the community.
Vague pushback rarely works. Tell AI exactly what it missed.
AI often reveals its own inconsistency when asked directly.
Every AI platform has thumbs down, flag, or feedback buttons. Use them. This is how models learn over time.
Post your comparisons. Show your community what you tested and what you got. This builds collective awareness and puts pressure on AI companies to do better.
Highest-impact actions you can take right now.
File a feedback report with AI companies. They track these. Name exactly what was missing or skewed.
Post your findings on social media, a blog, or your community newsletter.
If your school, synagogue, or company uses AI tools, they need to know about this.
Organizations doing this work need visibility and support.
Every action counts. Click to check off each step.
Get the latest prompts, share your findings, stay connected.
One sign-up. Pick what you want. No spam.
This runs on nothing but time and small contributions. See what your support funds or donate by card.
Every contribution keeps this campaign going.
Secure payment via Stripe - No account needed