OpenAI’s upcoming AI model built 948 working Linux root exploits in a single run

An au­to­ma­ted sys­tem run­ning on an up­co­ming AI model tur­ned hun­dreds of Linux ker­nel bugs into wor­king ex­ploits that give an or­di­nary user root ac­cess, and wrote a patch for each un­derl­ying flaw, Ope­nAI re­vea­led in a new re­search post. Sour­ce:Ope­nAI The com­pany’s Cyber team ran the pi­pe­li­ne in mid-Oc­to­ber 2026 against the la­test Linux ker­nel, with tar­get ma­chi­nes run­ning a stock desk­top ope­ra­ting sys­tem from a com­monly used dis­tri­bu­tion. The run ended with 948 … More → The post Ope­nAI’s up­co­ming AI model built 948 wor­king Linux root ex­ploits in a sin­gle run ap­pea­red first on Help Net Se­cu­rity .