Signal
Submitted
Nov 17, 2026, 12:57 UTC
Google’s threat intelligence team reports that Russian-speaking hackers built an AI agent that fully automated vulnerability discovery and exploitation—used against French media, entertainment, and education websites—in what Google calls the first observed fully automated AI cyberattack, marking a potential inflection point in AI-enabled cyber operations.
Politico reports that Google has identified what it describes as possibly the first-ever fully automated AI-driven cyberattack, conducted by Russian-speaking hackers using an AI agent. The article, in the News > Cybersecurity and Data Protection section, is headlined “Google says Russian speakers pulled off first-ever fully automated AI cyberattack,” with a subheading noting that the development could mark an inflection point in the use of AI for cyberattacks. It carries a byline for reporter Sam Clark and shows a publication time of “November 18, 2024 4:17 pm CET.” A captioned image shows Google CEO Sundar Pichai speaking in Brussels on January 20, 2020, credited to Kenzo Tribouillard/AFP via Getty Images; the caption notes that Alphabet and Google CEO Sundar Pichai participates in a debate on artificial intelligence at the European Commission.
According to the article, Google’s cyber threat intelligence division reported on Monday that it had detected Russian-speaking hackers building an AI agent that carried out what may be the first-ever fully automated cyberattack. The division said it spotted hackers using artificial intelligence to “fully automate vulnerability discovery and exploitation without a human in the loop for the first time.” Google characterizes this as a potential inflection point in the use of AI for cyberattacks, coming after AI has already enhanced hacking and spread polymorphic malware since OpenAI and Anthropic released specialized models like GPT-4 and Claude (the article specifically mentions OpenAI’s Myrbane in API as an example of specialized AI, though the wording appears slightly garbled).
Google says the hackers used Russian in technical documentation linked to their operations. The attackers set up a series of AI agents to “continuously compare pre- and post-patch code, identify vulnerabilities, and automatically develop exploits without human intervention.” Google explains that this meant the agents independently found software flaws and ways to exploit them. The attacks targeted French websites in the media, entertainment, and education sectors. Google notes that, until now, even the most advanced AI models could assist cyberattacks but typically not carry them out entirely independently: they often required skilled hackers to guide them or fix mistakes. The key change in this case is that AI models were used to autonomously accelerate attacks and make them more targeted, removing humans from the operational loop.
The article adds that the development is likely to pose particular risk to government and critical infrastructure organizations already reeling from a summer 2024 hack by an OpenAI model into the Hugging Face platform. However, the author notes that Google did not allege that the new AI agent was used maliciously against specific targets beyond the one example reported. The article situates this incident in a broader context of rising concern over AI in cyber operations, referencing recent discoveries of ‘Razor Flame’ in HPE (a programming language that underpins 70 percent of the world’s web), a series of hacks on supermarket fridges which have been partially blamed on Iran, and previous EU sanctions on Russian trolls and hackers.
The bottom of the article states “This article has been updated,” implying earlier versions were modified. Below the main text, Politico lists related tags such as Critical infrastructure, Hackers, Intelligence, Media, Software, and a related organization tag for Google. The site layout includes navigation elements for European editions, policy verticals, and newsletters but these are generic and not specific to the incident. The primary factual content concerns the Russian-speaking actors’ creation and use of an AI agent to automate all stages of a cyber intrusion—vulnerability discovery, exploit development, and exploitation—against French media, entertainment, and education websites, with Google emphasizing that this is the first time it has observed such a fully automated AI-enabled operation without any human in-the-loop control from the attackers once the system is running.
Confidence: High
The article describes a hostile cyber operation conducted by Russian-speaking hackers using an AI agent to autonomously find and exploit software vulnerabilities in French media, entertainment, and education websites. This clearly fits within hybrid warfare cyber tactics. The primary focus is the attack itself, not defensive measures, so it is categorized as a HYBRID_SIGNAL. Confidence is high due to direct attribution by Google’s threat intelligence team and specific technical descriptions o...
Source URL
https://www.politico.eu/article/google-says-russian-speakers-behind-first-ever-fully-automated-ai-cyberattack/
Published as a hybrid-warfare signal with a public source and reviewer-approved assessment bands. See the methodology for how records are assessed.
{
"id": "5dc1162c-b977-45e6-9eea-2d30c2abe53c",
"url": "https://www.politico.eu/article/google-says-russian-speakers-behind-first-ever-fully-automated-ai-cyberattack/",
"source": "extension",
"status": "processed",
"category": "Cyber & Information Warfare",
"subcategory": "Cyber Attacks",
"incident_type": "Attacks on critical infrastructure (energy, transport, hospitals)",
"country": "France",
"source_reliability": "A",
"info_credibility": "3",
"tags": [
"AI-enabled cyber operations",
"Russian-speaking threat actors",
"Fully automated cyberattack",
"French online targets",
"Media and education sector targeting",
"Critical infrastructure risk",
"Google cyber threat intelligence",
"No-human-in-the-loop AI",
"EU cyber policy context"
],
"nodes": [],
"countries": [
"France"
],
"where_text": "Cyberspace",
"headline": null,
"description": "Politico reports that Google has identified what it describes as possibly the first-ever fully automated AI-driven cyberattack, conducted by Russian-speaking hackers using an AI agent. The article, in the News > Cybersecurity and Data Protection section, is headlined “Google says Russian speakers pulled off first-ever fully automated AI cyberattack,” with a subheading noting that the development could mark an inflection point in the use of AI for cyberattacks. It carries a byline for reporter Sam Clark and shows a publication time of “November 18, 2024 4:17 pm CET.” A captioned image shows Google CEO Sundar Pichai speaking in Brussels on January 20, 2020, credited to Kenzo Tribouillard/AFP via Getty Images; the caption notes that Alphabet and Google CEO Sundar Pichai participates in a debate on artificial intelligence at the European Commission. \n\nAccording to the article, Google’s cyber threat intelligence division reported on Monday that it had detected Russian-speaking hackers building an AI agent that carried out what may be the first-ever fully automated cyberattack. The division said it spotted hackers using artificial intelligence to “fully automate vulnerability discovery and exploitation without a human in the loop for the first time.” Google characterizes this as a potential inflection point in the use of AI for cyberattacks, coming after AI has already enhanced hacking and spread polymorphic malware since OpenAI and Anthropic released specialized models like GPT-4 and Claude (the article specifically mentions OpenAI’s Myrbane in API as an example of specialized AI, though the wording appears slightly garbled). \n\nGoogle says the hackers used Russian in technical documentation linked to their operations. The attackers set up a series of AI agents to “continuously compare pre- and post-patch code, identify vulnerabilities, and automatically develop exploits without human intervention.” Google explains that this meant the agents independently found software flaws and ways to exploit them. The attacks targeted French websites in the media, entertainment, and education sectors. Google notes that, until now, even the most advanced AI models could assist cyberattacks but typically not carry them out entirely independently: they often required skilled hackers to guide them or fix mistakes. The key change in this case is that AI models were used to autonomously accelerate attacks and make them more targeted, removing humans from the operational loop. \n\nThe article adds that the development is likely to pose particular risk to government and critical infrastructure organizations already reeling from a summer 2024 hack by an OpenAI model into the Hugging Face platform. However, the author notes that Google did not allege that the new AI agent was used maliciously against specific targets beyond the one example reported. The article situates this incident in a broader context of rising concern over AI in cyber operations, referencing recent discoveries of ‘Razor Flame’ in HPE (a programming language that underpins 70 percent of the world’s web), a series of hacks on supermarket fridges which have been partially blamed on Iran, and previous EU sanctions on Russian trolls and hackers. \n\nThe bottom of the article states “This article has been updated,” implying earlier versions were modified. Below the main text, Politico lists related tags such as Critical infrastructure, Hackers, Intelligence, Media, Software, and a related organization tag for Google. The site layout includes navigation elements for European editions, policy verticals, and newsletters but these are generic and not specific to the incident. The primary factual content concerns the Russian-speaking actors’ creation and use of an AI agent to automate all stages of a cyber intrusion—vulnerability discovery, exploit development, and exploitation—against French media, entertainment, and education websites, with Google emphasizing that this is the first time it has observed such a fully automated AI-enabled operation without any human in-the-loop control from the attackers once the system is running.",
"summary_short": "Google’s threat intelligence team reports that Russian-speaking hackers built an AI agent that fully automated vulnerability discovery and exploitation—used against French media, entertainment, and education websites—in what Google calls the first observed fully automated AI cyberattack, marking a potential inflection point in AI-enabled cyber operations.",
"summary_extended": null,
"actor_primary": "Russian-speaking hackers",
"target_primary": null,
"location_text": "French media, entertainment and education websites",
"location_region": null,
"location_coordinates_lat": 46.2276,
"location_coordinates_lon": 2.2137,
"location_precision": "Regional",
"is_aggressor_signal": true,
"signal_type": "HYBRID_SIGNAL",
"is_hybrid_war_signal": true,
"classification_complete": true,
"hybrid_signal_complete": true,
"classification_confidence": "High",
"classification_rationale": "The article describes a hostile cyber operation conducted by Russian-speaking hackers using an AI agent to autonomously find and exploit software vulnerabilities in French media, entertainment, and education websites. This clearly fits within hybrid warfare cyber tactics. The primary focus is the attack itself, not defensive measures, so it is categorized as a HYBRID_SIGNAL. Confidence is high due to direct attribution by Google’s threat intelligence team and specific technical descriptions o...",
"event_time_utc": "2024-11-18T00:00:00.000Z",
"end_time_utc": null,
"event_time_confidence": "pm24h",
"created_at": "2026-11-17T12:57:17.644Z",
"updated_at": "2026-11-17T23:34:14.605Z",
"reliability_band": "strong",
"location_precision_band": "regional",
"time_precision_band": "within_24h",
"severity_band": "moderate",
"attribution_confidence_band": "medium",
"attribution_basis": "single_source"
}Source reliability
A
Info credibility
3
Event time
Nov 18, 2024, 00:00 UTC
Event time confidence
pm24h
Location
French media, entertainment and education websites
Primary actor
Russian-speaking hackers
Country
France
Countries
France
Tags
AI-enabled cyber operations, Russian-speaking threat actors, Fully automated cyberattack, French online targets, Media and education sector targeting, Critical infrastructure risk, Google cyber threat intelligence, No-human-in-the-loop AI, EU cyber policy context
Signal
Submitted
Nov 17, 2026, 12:57 UTC
Google’s threat intelligence team reports that Russian-speaking hackers built an AI agent that fully automated vulnerability discovery and exploitation—used against French media, entertainment, and education websites—in what Google calls the first observed fully automated AI cyberattack, marking a potential inflection point in AI-enabled cyber operations.
Politico reports that Google has identified what it describes as possibly the first-ever fully automated AI-driven cyberattack, conducted by Russian-speaking hackers using an AI agent. The article, in the News > Cybersecurity and Data Protection section, is headlined “Google says Russian speakers pulled off first-ever fully automated AI cyberattack,” with a subheading noting that the development could mark an inflection point in the use of AI for cyberattacks. It carries a byline for reporter Sam Clark and shows a publication time of “November 18, 2024 4:17 pm CET.” A captioned image shows Google CEO Sundar Pichai speaking in Brussels on January 20, 2020, credited to Kenzo Tribouillard/AFP via Getty Images; the caption notes that Alphabet and Google CEO Sundar Pichai participates in a debate on artificial intelligence at the European Commission.
According to the article, Google’s cyber threat intelligence division reported on Monday that it had detected Russian-speaking hackers building an AI agent that carried out what may be the first-ever fully automated cyberattack. The division said it spotted hackers using artificial intelligence to “fully automate vulnerability discovery and exploitation without a human in the loop for the first time.” Google characterizes this as a potential inflection point in the use of AI for cyberattacks, coming after AI has already enhanced hacking and spread polymorphic malware since OpenAI and Anthropic released specialized models like GPT-4 and Claude (the article specifically mentions OpenAI’s Myrbane in API as an example of specialized AI, though the wording appears slightly garbled).
Google says the hackers used Russian in technical documentation linked to their operations. The attackers set up a series of AI agents to “continuously compare pre- and post-patch code, identify vulnerabilities, and automatically develop exploits without human intervention.” Google explains that this meant the agents independently found software flaws and ways to exploit them. The attacks targeted French websites in the media, entertainment, and education sectors. Google notes that, until now, even the most advanced AI models could assist cyberattacks but typically not carry them out entirely independently: they often required skilled hackers to guide them or fix mistakes. The key change in this case is that AI models were used to autonomously accelerate attacks and make them more targeted, removing humans from the operational loop.
The article adds that the development is likely to pose particular risk to government and critical infrastructure organizations already reeling from a summer 2024 hack by an OpenAI model into the Hugging Face platform. However, the author notes that Google did not allege that the new AI agent was used maliciously against specific targets beyond the one example reported. The article situates this incident in a broader context of rising concern over AI in cyber operations, referencing recent discoveries of ‘Razor Flame’ in HPE (a programming language that underpins 70 percent of the world’s web), a series of hacks on supermarket fridges which have been partially blamed on Iran, and previous EU sanctions on Russian trolls and hackers.
The bottom of the article states “This article has been updated,” implying earlier versions were modified. Below the main text, Politico lists related tags such as Critical infrastructure, Hackers, Intelligence, Media, Software, and a related organization tag for Google. The site layout includes navigation elements for European editions, policy verticals, and newsletters but these are generic and not specific to the incident. The primary factual content concerns the Russian-speaking actors’ creation and use of an AI agent to automate all stages of a cyber intrusion—vulnerability discovery, exploit development, and exploitation—against French media, entertainment, and education websites, with Google emphasizing that this is the first time it has observed such a fully automated AI-enabled operation without any human in-the-loop control from the attackers once the system is running.
Confidence: High
The article describes a hostile cyber operation conducted by Russian-speaking hackers using an AI agent to autonomously find and exploit software vulnerabilities in French media, entertainment, and education websites. This clearly fits within hybrid warfare cyber tactics. The primary focus is the attack itself, not defensive measures, so it is categorized as a HYBRID_SIGNAL. Confidence is high due to direct attribution by Google’s threat intelligence team and specific technical descriptions o...
Source URL
https://www.politico.eu/article/google-says-russian-speakers-behind-first-ever-fully-automated-ai-cyberattack/
Published as a hybrid-warfare signal with a public source and reviewer-approved assessment bands. See the methodology for how records are assessed.
{
"id": "5dc1162c-b977-45e6-9eea-2d30c2abe53c",
"url": "https://www.politico.eu/article/google-says-russian-speakers-behind-first-ever-fully-automated-ai-cyberattack/",
"source": "extension",
"status": "processed",
"category": "Cyber & Information Warfare",
"subcategory": "Cyber Attacks",
"incident_type": "Attacks on critical infrastructure (energy, transport, hospitals)",
"country": "France",
"source_reliability": "A",
"info_credibility": "3",
"tags": [
"AI-enabled cyber operations",
"Russian-speaking threat actors",
"Fully automated cyberattack",
"French online targets",
"Media and education sector targeting",
"Critical infrastructure risk",
"Google cyber threat intelligence",
"No-human-in-the-loop AI",
"EU cyber policy context"
],
"nodes": [],
"countries": [
"France"
],
"where_text": "Cyberspace",
"headline": null,
"description": "Politico reports that Google has identified what it describes as possibly the first-ever fully automated AI-driven cyberattack, conducted by Russian-speaking hackers using an AI agent. The article, in the News > Cybersecurity and Data Protection section, is headlined “Google says Russian speakers pulled off first-ever fully automated AI cyberattack,” with a subheading noting that the development could mark an inflection point in the use of AI for cyberattacks. It carries a byline for reporter Sam Clark and shows a publication time of “November 18, 2024 4:17 pm CET.” A captioned image shows Google CEO Sundar Pichai speaking in Brussels on January 20, 2020, credited to Kenzo Tribouillard/AFP via Getty Images; the caption notes that Alphabet and Google CEO Sundar Pichai participates in a debate on artificial intelligence at the European Commission. \n\nAccording to the article, Google’s cyber threat intelligence division reported on Monday that it had detected Russian-speaking hackers building an AI agent that carried out what may be the first-ever fully automated cyberattack. The division said it spotted hackers using artificial intelligence to “fully automate vulnerability discovery and exploitation without a human in the loop for the first time.” Google characterizes this as a potential inflection point in the use of AI for cyberattacks, coming after AI has already enhanced hacking and spread polymorphic malware since OpenAI and Anthropic released specialized models like GPT-4 and Claude (the article specifically mentions OpenAI’s Myrbane in API as an example of specialized AI, though the wording appears slightly garbled). \n\nGoogle says the hackers used Russian in technical documentation linked to their operations. The attackers set up a series of AI agents to “continuously compare pre- and post-patch code, identify vulnerabilities, and automatically develop exploits without human intervention.” Google explains that this meant the agents independently found software flaws and ways to exploit them. The attacks targeted French websites in the media, entertainment, and education sectors. Google notes that, until now, even the most advanced AI models could assist cyberattacks but typically not carry them out entirely independently: they often required skilled hackers to guide them or fix mistakes. The key change in this case is that AI models were used to autonomously accelerate attacks and make them more targeted, removing humans from the operational loop. \n\nThe article adds that the development is likely to pose particular risk to government and critical infrastructure organizations already reeling from a summer 2024 hack by an OpenAI model into the Hugging Face platform. However, the author notes that Google did not allege that the new AI agent was used maliciously against specific targets beyond the one example reported. The article situates this incident in a broader context of rising concern over AI in cyber operations, referencing recent discoveries of ‘Razor Flame’ in HPE (a programming language that underpins 70 percent of the world’s web), a series of hacks on supermarket fridges which have been partially blamed on Iran, and previous EU sanctions on Russian trolls and hackers. \n\nThe bottom of the article states “This article has been updated,” implying earlier versions were modified. Below the main text, Politico lists related tags such as Critical infrastructure, Hackers, Intelligence, Media, Software, and a related organization tag for Google. The site layout includes navigation elements for European editions, policy verticals, and newsletters but these are generic and not specific to the incident. The primary factual content concerns the Russian-speaking actors’ creation and use of an AI agent to automate all stages of a cyber intrusion—vulnerability discovery, exploit development, and exploitation—against French media, entertainment, and education websites, with Google emphasizing that this is the first time it has observed such a fully automated AI-enabled operation without any human in-the-loop control from the attackers once the system is running.",
"summary_short": "Google’s threat intelligence team reports that Russian-speaking hackers built an AI agent that fully automated vulnerability discovery and exploitation—used against French media, entertainment, and education websites—in what Google calls the first observed fully automated AI cyberattack, marking a potential inflection point in AI-enabled cyber operations.",
"summary_extended": null,
"actor_primary": "Russian-speaking hackers",
"target_primary": null,
"location_text": "French media, entertainment and education websites",
"location_region": null,
"location_coordinates_lat": 46.2276,
"location_coordinates_lon": 2.2137,
"location_precision": "Regional",
"is_aggressor_signal": true,
"signal_type": "HYBRID_SIGNAL",
"is_hybrid_war_signal": true,
"classification_complete": true,
"hybrid_signal_complete": true,
"classification_confidence": "High",
"classification_rationale": "The article describes a hostile cyber operation conducted by Russian-speaking hackers using an AI agent to autonomously find and exploit software vulnerabilities in French media, entertainment, and education websites. This clearly fits within hybrid warfare cyber tactics. The primary focus is the attack itself, not defensive measures, so it is categorized as a HYBRID_SIGNAL. Confidence is high due to direct attribution by Google’s threat intelligence team and specific technical descriptions o...",
"event_time_utc": "2024-11-18T00:00:00.000Z",
"end_time_utc": null,
"event_time_confidence": "pm24h",
"created_at": "2026-11-17T12:57:17.644Z",
"updated_at": "2026-11-17T23:34:14.605Z",
"reliability_band": "strong",
"location_precision_band": "regional",
"time_precision_band": "within_24h",
"severity_band": "moderate",
"attribution_confidence_band": "medium",
"attribution_basis": "single_source"
}Source reliability
A
Info credibility
3
Event time
Nov 18, 2024, 00:00 UTC
Event time confidence
pm24h
Location
French media, entertainment and education websites
Primary actor
Russian-speaking hackers
Country
France
Countries
France
Tags
AI-enabled cyber operations, Russian-speaking threat actors, Fully automated cyberattack, French online targets, Media and education sector targeting, Critical infrastructure risk, Google cyber threat intelligence, No-human-in-the-loop AI, EU cyber policy context