@middleeasteye: Forget influencing people - Israel’s latest play is influencing what AI tells them. We got the tour. @blindeyerick

Tan.Jung :

Yes. Coz it refuses to describe Israeli genocide in Gaza.

Hypernormalized :

They do this with Wikipedia as well.

Lola_Jupiter :

Our whole world is spiraling into chaos and AI is speeding up the process

cunnnnnnnmnnmnnnnnnmmnnnnnnt :

It occurs on an even deeper level than this. The RLHF post-training stage is able to be geared towards whatever the fuck the people in charge want it to (as well as setting up guard rails preventing the model from discussing anything they don’t want it to) . they make fun of China’s models like Deepseek because it’s hesitant to reference Tiananmen Square but often it will. But just try and get Google Gemini to discuss Gaza and whether it’s genocide😂. If it was only due to disproportionate training material (which would exist anyway even if they didn’t try to game it) it would typically concede upon presentation of new, reliable information (eg all the institutions who have said it is a genocide) but instead it either flat out rejects, “gaslights”, or literally just shuts down and isn’t able to reply. No surprise given Sam Altman is Jewish and has plenty of Zionist ties, Google is openly pro Israel/zionism/genocide, Anthropic is owned/founded and ran by Dario Amodei who is Jewish but tbh he seems to be far more ethical than the others (not that that’s saying much) and is definitely more intelligent (than other founders except Demis Hassabis) and I don’t have enough experience with Claude to know how it responds to talks of Palestine. Anyway, the point is, please push your government (esp if you’re American) to call on these companies to be more transparent about the training process otherwise there is nothing stopping them steering the model toward whatever a single person or handful of very rich people want it to tell you. I say this as someone studying and working directly in AI research (at a uni though and not directly making foundational models, just the underlying architectures/math).

Vault Dweller :

They want to own everything. Now its AI.

anon1234567 :

every time someone tells me they're using chatgpt to answer basic questions. it's programmable people. that makes it a tool of control.

The Dems Suck Podcast :

who is this guy bc he is very good.

shaneanthony57 :

No doubt they are trying , they try and distort everything

notfluffy :

yes

amy :

The world will be safer and flourishes and peace without Israel and USA

Ellie :

They are everywhere

ASYA :

Elllee.bs 🇱🇧 :

GOD BLESS ISRAEL

Moody :

Ahahaha horrifying report, but absolutely brilliant job by telling it @BlindEyeRick 😂👏👏👏

YT studios :

😏😏😏

Freedom and Humanity :

😡😡😡

BlazeDog11 :

How could he have 3 million followers and only a few comments? Jews are always involved