Grok, ChatGPT, Claude and Gemini answer who would win if one AI hacked another

TechnologyInternationalEl ColombianoNov 11, 2026, 9:01 AMAI summary

El Colombiano asked four language models —Grok (xAI), Claude (Anthropic), ChatGPT (OpenAI) and Gemini (Google)— what would happen if one AI hacked another. Grok said it would have a "good chance" because it is less restricted, but admitted that "nobody is invincible." Claude and ChatGPT agreed that no model is an autonomous hacker and that the outcome would depend on security architecture and permissions. Gemini said the real contest would be between the companies' cybersecurity teams and infrastructure.

In 45 seconds

Grok (xAI) said it would have a "good chance" because it is less restricted, though it argued that "nobody is invincible."
Claude (Anthropic) said hacking an AI does not work like in the movies and that the best-configured and most isolated system would win, not the "smartest."
ChatGPT (OpenAI) placed Claude and itself on the front line of a cybersecurity contest, with Gemini close behind.

Read the full story at El Colombiano

IASummary generated by artificial intelligence from a story by El Colombiano. It may contain errors: the original source prevails.

inteligencia artificial
ciberseguridad
OpenAI
Anthropic
Google

Comments

Be respectful. Comments reported by several people are hidden until a moderator reviews them.

You do not need an account to read, only to comment. Sign in with a code sent to your email or with your Google account. We only show the name you choose, never your email.

Keep reading

Anthropic unveils Claude Fable 5.2 and Mythos 5.2, its new AI models
Anthropic presented two new models in its Claude family: Fable 5.2 and Mythos 5.2, with improvements in reasoning, programming and visual recognition. According to the company, its engineers used them to complete software migrations previously considered unfeasible and reported progress in research on black hole physics and robotics, as well as the formalization of a mathematical proof related to Fermat's Last Theorem in 30 hours. Fable 5.2 is available to the general public; Mythos 5.2 will have restricted distribution to organizations that pass verification processes.
TechnologyInternationalRevista Semana11:00 AMAI
Cybersecurity expert: social media self-regulation "has not delivered results"
Jorge Flores, director of PantallasAmigas, an organization that promotes safe and healthy use of the internet and social media, said on 6AM W that platform self-regulation "has not delivered results" because insufficient means are being used to fight cybercrime, especially sexual abuse of minors. He noted that directives exist in Europe and the United States seeking to impose a stricter, more transparent and punishable framework on platforms, and said shutting them down in a country would be neither very possible nor very effective.
TechnologyInternationalCaracol Radio9:58 AMAI
CoreWeave CEO cancels share sale program after beating revenue expectations
CoreWeave CEO Michael Intrator told CNBC he canceled his 10b5-1 stock sale plan after the company's third-quarter results, in which it beat estimates with 168% year-over-year revenue growth and raised its 2026 forecast. Under that plan, announced in November 2025, he had sold 8.26 million shares at an average price of $96.14, for $794.2 million. Intrator said he does not anticipate any impact from the regulatory shift following the Democratic electoral victory.
TechnologyInternationalCNBC9:20 AMAI
TransMilenio adds 10 fully electric articulated buses, the first of their kind in Bogotá
TransMilenio added ten fully electric articulated buses, the first of this type to operate in Bogotá, each with capacity for 160 passengers. The vehicles were manufactured by Yutong in China and fitted with Intelligent Transport Systems installed by LG. At Patio Sur, operated by Consorcio Express, a first charging island was prepared; work on the remaining islands will allow up to 116 electric vehicles to operate, with a final fleet of 74 articulated and 42 bi-articulated buses.
TechnologyColombiaPulzo8:06 AMAI

Grok, ChatGPT, Claude and Gemini answer who would win if one AI hacked another

In 45 seconds

Comments

Keep reading

Anthropic unveils Claude Fable 5.2 and Mythos 5.2, its new AI models

Cybersecurity expert: social media self-regulation "has not delivered results"

CoreWeave CEO cancels share sale program after beating revenue expectations

TransMilenio adds 10 fully electric articulated buses, the first of their kind in Bogotá