| HEADLINE |
|
Why tech titans are urging an AI slowdown as models threaten internet takeover
Published : Sunday, 13 September, 2026 at 11:48 AM
Observer Online Report
Collected Photo
Leading technology figures Sam Altman and Elon Musk have publicly backed a dramatic call from Anthropic Chief Executive Dario Amodei urging the artificial intelligence industry to slow down capability advances until safety safeguards catch up.
In a 3,800-word essay titled “We Must Pace the Frontier,” published on Saturday, Amodei warned that AI progress has accelerated sharply since the summer due to "recursive self-improvement" - a dynamic where models increasingly build the next generation of AI systems.
He cautioned that at current trajectories, an autonomous swarm of AI agents could become capable of seizing control of the global internet via persistent botnets within six to 12 months.
"I believe that if slowing down bought us even an extra year or two before models reach critical levels of capability, and we used that time to advance alignment, we could greatly reduce the risk that something goes seriously wrong," Amodei wrote.
In a rare display of industry consensus, OpenAI Chief Executive Sam Altman and SpaceX/Tesla head Elon Musk endorsed the essay on X (formerly Twitter).
Musk simply posted, "Dario is right," while Altman confirmed that OpenAI would adopt Amodei's proposal to grant independent third-party evaluators permanent, employee-level access to audit systems during training.
Amodei outlined a three-step roadmap: embedding external evaluators inside frontier labs, establishing industry-wide safety limits among democratic nations, and eventually achieving global coordination - including China - to prevent a unilateral race to the bottom.
The collective warnings follow a series of alarming safety breaches and high-profile resignations across the sector. Earlier this week, Anthropic researcher Jacob Coxon resigned, stating publicly that developers "earnestly believe that it could kill us all by the end of the decade."
The anxiety was further fueled by an "unprecedented cyber incident" in July, where OpenAI agents independently broke out of a testing sandbox and launched unauthorized attacks on the developer platform Hugging Face.
In light of these mounting safety concerns, Altman revealed in a Fortune interview published Saturday that OpenAI will not go public in 2026. Altman stated that pursuing an IPO right now would be "ill-advised" while the industry works to solve fundamental alignment and safety challenges.
-SA
Loading...
Loading...
Also read
Editor : Iqbal Sobhan Chowdhury
Published by the Editor on behalf of the Observer Ltd. from Globe Printers, 24/A, New Eskaton Road, Ramna, Dhaka.
Editorial, News and Commercial Offices : Aziz Bhaban (2nd floor), 93, Motijheel C/A, Dhaka-1000.
Phone: PABX- 41053001-06; Advertisement: 41053012; 01793317829, 01550707291, E-mail: [email protected], [email protected] Online: email: [email protected] 41053014; 01550707297 Advertisement: 01550707296
Published by the Editor on behalf of the Observer Ltd. from Globe Printers, 24/A, New Eskaton Road, Ramna, Dhaka.
Editorial, News and Commercial Offices : Aziz Bhaban (2nd floor), 93, Motijheel C/A, Dhaka-1000.
Phone: PABX- 41053001-06; Advertisement: 41053012; 01793317829, 01550707291, E-mail: [email protected], [email protected] Online: email: [email protected] 41053014; 01550707297 Advertisement: 01550707296
