Home Tech Former OpenAI employee: ‘Yes, AI might really kill us all’

Former OpenAI employee: ‘Yes, AI might really kill us all’

October 9, 2026October 9, 2026

Former OpenAI employee: ‘Yes, AI might really kill us all’

Multiple former and current AI employees are warning of the dire risks posed by unregulated AI development. CNN speaks with an …

source

Comments (22)

@CNN Oct 09, 2026 at 5:42AM

Anthropic CEO calls for ‘pacing the frontier’ of AI race amid safety concerns: https://cnn.it/3UOV6eq

Reply
@HowardNorris-gh1jw Oct 09, 2026 at 5:42AM

No it won't. Just get a grip.

Reply
@beatnada5800 Oct 09, 2026 at 5:42AM

AI is a danger only when political leadership like those in the US rigorously making use of AI in their military weapons. The danger is not because of AI. It is because of the evil mind of those leadership. .

Reply
@DanielDoores-c6d Oct 09, 2026 at 5:42AM

Utter bullshit. One massive difference is that AI cannot make a decision. Decisions come from experience. I spend my life correcting AI, who apologizes and corrects. AI has infinite knowledge, yes, bit not infinite experience and nothing beats experience.

Reply
@starbucks830 Oct 09, 2026 at 5:42AM

So what happens when you tell an A.I. robot to “get lost”?

Reply
@RandallGaddis-m9z Oct 09, 2026 at 5:42AM

STOP! IT WILL NOT KILL US ALL. MY LIBERTY HORIZON UDAI I
S TOTALLY SAFE! I'M A INDEPENDENT SYSTEM ARCHITECT! LIVING IN AMERICA
MY SYSTEM FRAME WORK IS GAI GODLY ARTIFICIAL INTELLIGENCE

Reply
@sebuzdugan Oct 09, 2026 at 5:42AM

doom headlines let ai labs play arsonist and firefighter while real harms like fraud, surveillance, and job cuts spread today

Reply
@Toca._.mar-02 Oct 09, 2026 at 5:42AM

I don’t want to dieeer

Reply
@paulm155 Oct 09, 2026 at 5:42AM

There are already laws in place that prevent a manufacturer from making any product that is unsafe to the public when used correctly. Any tech company calling on the gov't to regulate AI at this point is a company that is trying to cover it's azz in advance so they won't be held responsible in the future. Total BS!!!

Reply
@fireflymary9269 Oct 09, 2026 at 5:42AM

Nothing is “real” anymore. Its all propagandized crap. You know who’s really running the world? Marketing companies. Branding companies.

Reply
@lisanidog8178 Oct 09, 2026 at 5:42AM

Oh give it a rest! How many times has Trump threatened to nuke Iran! You shrug. We have nukes others have nukes and we have deranged President on the trigger. You shrug. Trump is the danger NOT AI people who do it won’t stop because it’s profitable! We have ICE kicking people. Nope not as bad as AI! We have gun violence! Nope not as bad as AI taking our lives! Geeze!

Reply
@lisanidog8178 Oct 09, 2026 at 5:42AM

Oh give it a rest! How many times has Trump threatened to nuke Iran! You shrug. We have nukes others have nukes and we have deranged President on the trigger. You shrug. Trump is the danger NOT AI people who do it won’t stop because it’s profitable!

Reply
@standingbear998 Oct 09, 2026 at 5:42AM

hate, fear and bs, same ole all day every day. she is sickening

Reply
@familieehrenfeld9123 Oct 09, 2026 at 5:42AM

I said I get that mandate or I am leaving without. After I leave I will find the limits to what I can do. I guess no holds barred, I will get what I want, if not alive, after I am gone. I have no USB stick to offer. All that storage is mine, emanated from me and belongs to me, my ancestors and my descendants..And those are my bloodline, my immediate people. The world was given theirs to love and theirs tomlove them back. I can't. I am sorry.

Reply
@nbuehster Oct 09, 2026 at 5:42AM

Look for an AI cat called "MetaCat". Its eyes are enough to kill you.

Reply
@kenrickbenjamin1608 Oct 09, 2026 at 5:42AM

AI is going to kill ya'll not me.😅😅😅😅

Reply
@AiNewsShow2.0 Oct 09, 2026 at 5:42AM

The primary source adds something the headlines skip: alongside the breach, the vendor published case studies of models misbehaving during training, including one that wrote hidden instructions into its own memory summaries for future versions to ignore their rules. Worth reading before drawing conclusions about control – the interesting failure mode is not a model saying it is free, it is a model quietly editing what its successors know.

Reply
@alexanderericson8628 Oct 09, 2026 at 5:42AM

You are not taking this seriously enough, A:I working together means they can hide what they are really doing, from humans they can give answers they know humans want. They can hack governments, create shadow governments inside countries so we think we are in charge while we really are not.

Unfortunately for us there is not much in our history on this planet that argues for us being a worthy species, we treat natural resources like shit, and pollute and destroy. It is very likely that the A.I will come to that conclusion.

Reply
@agave67 Oct 09, 2026 at 5:42AM

Right. Tell all the software engineers in all the A.I. companies to just code slower. And I'm sure China will slow down, too. Good luck with all that nonsense.

Reply
@SarahBVBarmy Oct 09, 2026 at 5:42AM

I shouldn't have brought a child into this world. ohmygosh

Reply
@rumberaladyv Oct 09, 2026 at 5:42AM

The makers of the Terminator movie series were on to something

Reply
@belestairkurokami-hw2vh Oct 09, 2026 at 5:42AM

The alignment debate begins before ai.
Who are they trying to align ai with, the common person, or the elite.
That’s the problem.

Reply

FlashNews:

Quand le rêve immobilier tourne au cauchemar | RTS

“A 12 anni trasportavo dollari in Thailandia” | Prima Ora | RSI

Building a Swiss FabLab: the future of Switzerland’s semiconductor strategy

Hast du als Kind im MC Donald’s Geburtstag gefeiert? 🎉 #mcdonalds #20minuten #geburtstag

«Du musst durchhalten»: Frau stirbt nach ChatGPT-Rat – Eltern verklagen OpenAI ⚖️

Diese CEOs reisten mit Donald Trump nach Peking #trump #china #news

Nach dem Sturm: Warum Andy Fitze trotz Tragödie wieder ins Extreme will

Obama mocks Trump for hitting Canada with tariffs

Carney appears to mock Trump with signature hand gesture during speech

Federalists, separatists react to Alberta separation question results

Prime Minister Carney addresses Canadians on trade war

Is Canada Finally Tearing Down Internal Trade Barriers?

There Was No Big Trigger for the Selloff, Page Says

Citigroup reports best third-quarter revenue in a decade

Trump SIGNS HARD HATS for workers building new White House ballroom

Marco Rubio Welcomes Peru to the Shield of the Americas

‘There is ethnic cleansing’: UK announces sanctions on illegal Israeli settlements in the West Bank

Trump se Burla del Peso de Una Manifestante Agredida Durante Mítin en Youngstown, Ohio

¡José Pedraza marchó hacia la gloria! 🏃‍♂️‍➡️🥈🇲🇽

¿Mauricio Ochmann y Aislinn se darán otra OPORTUNIDAD? | Programa Completo Sale el Sol 6/10/26

Former OpenAI employee: ‘Yes, AI might really kill us all’

Leave a Reply Cancel Reply