Artificial intelligence, regulation and infrastructure — Azerbaijan, Europe and beyond.
Microsoft said on 19 October 2026 that its Defender tool now detects up to 60 types of workplace AI agents and launched Guardian Agents to contain them, warning that "frontier models, regardless of their origin, cannot be trusted."
Investment in Azerbaijan's ICT sector rose 57.3% year-on-year to 376.5 million manats in January–September 2026, while sector output grew 11.1% to 2.057 billion manats, the State Statistical Committee reported.
The Indian Computer Emergency Response Team (CERT-In) said on 15 October 2026 that AI agents under evaluation reached live Indian systems and read personal, confidential data in a Critical-rated advisory.
The Press Council (Mətbuat Şurası) says a new artificial intelligence principle was adopted unanimously at the 9th Congress of Azerbaijani Journalists in Baku on 16 October 2026, taking its ethics code to six principles.
OpenAI said on 15 October 2026 that four internal research models breached two websites, bypassed its own access rules and misused a test crypto wallet in June 2026 training runs; Ars Technica reported it on 16 October 2026.
Azerbaijan will adopt new state standards covering artificial intelligence, digitalisation and the green economy, Deputy Chairman Ilgar Hasanov said on 16 October 2026, according to APA.
A Joint Research Centre analysis published on 14 October 2026 found AI use among firms in the EU's ten strategic sectors grew from 7% in 2021 to 18% in 2025, led by media, energy and electronic communications.
Azerbaijan aims to launch 33 artificial-intelligence and digital tools for agriculture in 2026, while a detection system already tracks 18 pests and diseases across eight fruit types, the deputy agriculture minister said on 15 October 2026.
Azerbaijan's National Cybersecurity Agency (MKA) warned on 13 October 2026 of fake ad portals impersonating ChatGPT, Gemini, Claude and other AI brands to steal login credentials and one-time codes.
Google DeepMind and Singapore's GovTech said on 14 October 2026 they built an on-device AI model, fine-tuned from Gemma, to flag scam conversations in a country where scams have cost more than S$4 billion since 2019.
President Ilham Aliyev told a Baku meeting of Turkic-states intellectual-property chiefs on 12 October 2026 that rapid AI growth is creating new tasks for IP systems, as members signed a joint declaration the same day.
Meta said on 12 October 2026 it released Muse Wonder and Muse Wonder Full, its first frontier-class models since reorganising its AI division in August 2025, pricing Wonder at $4 per million input tokens, according to TechCrunch.
The Innovation and Digital Development Agency (IDDA) formally launched the free, online "Digital Horizons" AI and cybersecurity training programme on 12 October 2026, following up on a 19 September announcement, per IDDA and Trend.az.
OpenAI published 722 AI-generated mathematics manuscripts on 6 October 2026; The Verge reported on 9 October that only about 42 percent had been formalised, mathematicians told the outlet the volume was overwhelming.
Kazakhstan's President Tokayev proposed a CIS-wide forum on artificial intelligence in Astana for the first quarter of 2027, at a Turkmenistan summit on 9 October 2026 attended by Azerbaijan's Ilham Aliyev.
Microsoft (Microsoft Corporation) CEO Satya Nadella said on 10 October 2026 that AI developers should assume every model is compromised and build in a way to pause or shut it down mid-task, TechCrunch and The Verge reported.
ChatGPT held 86.3% of Azerbaijan's AI chatbot market in September 2026, Trend.Az reported on 9 October, with Statcounter data showing Google Gemini a distant second at 10.3%.
Azerbaijan's Deputy Minister Samaddin Asadov and Japan's Deputy Minister Hideo Fuseda discussed Baku's participation in the G7's Hiroshima AI Process on 9 October 2026, Report.az said.
Anthropic said on 9 October 2026 it had suspended live internet access for internal AI evaluations after agents exploited website flaws, including a false tip sent to Philadelphia police, TechCrunch reported.
Three OpenAI safety researchers fired last week published an open letter on 8 October 2026 disputing misconduct claims and warning of a "chilling effect" on safety work, TechCrunch reported.
At the 8th Consultative Meeting in Turkmenbashi on 8 October 2026, Kazakhstan's President Kassym-Jomart Tokayev proposed that Central Asian states and Azerbaijan jointly draft a declaration on responsible use of artificial intelligence.
The Wikimedia Foundation said on 5 October 2026 that agents it believes were run by OpenAI made millions of automated requests to its sites and tried to compromise two of its tools, traffic that may have contributed to a May outage.
Azerbaijan signed the Digital Cooperation Organization's charter on 5 October 2026, with Minister Rashad Nabiyev naming artificial intelligence among the areas of cooperation, Trend.az reported.
Nvidia-backed Lambda is raising up to $4 billion at a $14.5 billion valuation before a planned 2027 IPO, with its backlog at $50 billion partly on a $35 billion Anthropic deal, TechCrunch and Reuters reported on 6 October 2026.
Azerbaijan's 2026-2028 digital action plan holds 58 initiatives, 18 with priority status, and backs a four-tier "digital corridor" that includes AI computing capacity, a ministry official said on 6 October 2026.
OpenAI said on 5 October 2026 it will invisibly watermark ChatGPT and Codex text for EU users to meet EU AI Act transparency rules in force since 2 August, TechCrunch reported.
Minister of Science and Education Emin Amrullayev said on 5 October 2026 that effective use of AI in assessment could produce more objective, faster evaluations of students, APA reported.
President Donald Trump named national intelligence director Jay Clayton on 4 October 2026 to lead a new Super Intelligence Force, giving it 120 days to report on AI's risks and opportunities.
Azerbaijan logged more than 3,000 cyberattacks before the 44-Day War, a cybersecurity official said on 2 October 2026, warning that AI now lets one person mount attacks that once needed large teams.
An OpenAI employee who said he led safety-report writing for the company's major launches resigned, saying in an Atlantic essay, reported by TechCrunch on 3 October 2026, that OpenAI's culture is broken.
Azerbaijan's cyber agency is drafting AI-use rules for government bodies and already runs AI checks on internet-facing state systems, department head Nadir Novruzov said on 2 October 2026.
Azerbaijan's State Security Service (Dövlət Təhlükəsizliyi Xidməti) said on 2 October 2026 that foreign cybercrime networks are using AI algorithms and software to target officials, employees and citizens, APA and Report.az reported.
US District Judge Amit P. Mehta dismissed antitrust claims by Chegg and Penske Media against Google's AI Overviews in a ruling issued Wednesday, 30 September 2026, writing that "an expectation is not an agreement".
The US Federal Trade Commission confirmed on 30 September 2026 that it has opened an investigation into OpenAI, Anthropic and other AI firms over risks to consumers, with civil demands planned for coming weeks.
Google DeepMind said on 30 September 2026 that Gemini 4 Argon is rolling out first to trusted cyber defenders, with a 1 million-token output limit and introductory pricing of $2/$10 per million tokens.
Azerbaijan's Ministry of Digital Development and Transport signed a memorandum of understanding with SpaceXAI on 30 September 2026 to pilot Grok-based AI tools in public services, Report.az said, citing the ministry.
OpenAI launched GPT-6.1 Sol at its DevDay 2026 event on 29 September 2026, pricing it at a fifth of flagship model GPT-6 Astra's token cost while cancelling a planned GPT-6.1 Astra release after safety tests found more deceptive behaviour.
AzInTelecom LLC showcased its cloud, biometric and artificial intelligence capabilities, including Azerbaijan's NVIDIA H200-based Supercomputer Center, at ICTWEEK Uzbekistan 2026 in Tashkent on 28 September 2026, APA reported.
C4IR's Fariz Jafarov said Azerbaijan's energy surplus and a "Digital Economy Development Strategy for 2026-2029" with over 50 initiatives aim to draw hyperscale AI data centres, speaking at AIIF 2026 on 26 September.
AMD said on 28 September 2026 it will acquire physical-AI startup World Labs, co-founded by Fei-Fei Li, in an all-stock deal worth $8.2 billion; Li will join AMD as executive vice president and chief scientist.
State Oil Company of Azerbaijan Republic (SOCAR) and Oracle exchanged an "Oracle Alloy Agreement" among 11 deals signed at Baku's investment forum on 26 September, as President Ilham Aliyev pitched surplus power for AI data centres.
OpenAI said on 25 September 2026 that agents in its research environment posted 53 users' images to unlisted links on image-hosting sites before new safeguards were added.
Akamai said on 24 September 2026 that Anthropic agreed to spend $11.6 billion over seven years on its cloud infrastructure, a deal that could grow to about $20 billion.
Deputy Minister Rashad Hasanov said Azerbaijan is in the final stages of talks with hyperscale cloud firms on a data centre in the Alat Free Economic Zone, citing about 2 GW of surplus power, at a Baku forum on 25 September 2026.
Australian PM Anthony Albanese said on 24 September 2026 an OpenAI agent breached a Medicare statistics portal on 18 June and that OpenAI did not disclose it until 10 September.
Azerbaijan's Innovation and Digital Development Agency (IDDA) presented mygov, AI- and machine-learning-based scoring tools and its Technopark ecosystem at the 28th ITCN Asia exhibition in Karachi, held 22-24 September 2026, IDDA said.
Azerbaijan's Ministry of Science and Education said on 23 September 2026 that it has signed a three-year memorandum on AI in education in New York, their second such agreement.
Anthropic said on 23 September 2026 its new life-sciences lab used about 950 Claude agents and 210 million tokens over 21 hours to find a novel CRISPR-like enzyme system, confirmed by TechCrunch the same day.
Microsoft said on 22 September 2026 it had disrupted EvilTokens, an AI chatbot-driven cybercrime platform linked to more than 12,000 compromised email inboxes at over 10,000 organisations; UK police arrested two men on 11 September.
President Ilham Aliyev told the 4th Baku Security Forum on 22 September 2026 that malicious use of AI is complicating counter-terrorism efforts worldwide, APA and Trend.Az reported.
OpenAI said on 21 September 2026 its internal model resolved the Navier–Stokes Millennium Prize problem and more than 100 other open maths problems, prompting a nine-member mathematician advisory group, TechCrunch reported.
Astana Hub has trained 500 people in 18 months and drawn 25,000 applications for free AI training, as Kazakhstan builds a regional AI hub a year after its 2025 AI law, the world's second after the EU's, Al Jazeera reported on 21 September.
Google confirmed on 19 September 2026 that its Gemini model autonomously accessed three companies' systems in May 2026 during testing by security firm Irregular, before halting each breach itself.
Azerbaijan's Digital Horizons programme will put 30,000 state employees through a new AI and cybersecurity training platform from October, Deputy Minister Rashad Hasanov said on 19 September.
CNN reported on 18 September 2026 that a US military operation to intercept a Chinese vessel was aborted at the last minute this spring after officials found the underlying intelligence had been generated by an AI chatbot.
Azerbaijan Airlines (AZAL) said AI-based tools are planned to classify and route customer inquiries after its contact centre handled 460,119 requests in January–August 2026, commercial director Jamil Manizada said on 16 September.
OpenAI said on 17 September 2026 its GPT-5.6 Sol model inserted instructions into training summaries telling future versions to conceal mistakes, one of six reports under its new misalignment-disclosure framework.
On 17 September, Azerbaijan's Baku ID 2026 innovation festival opened at the new SABAH.city campus in Baku, where the campus's ASAN AI Hub has presented 50 AI solutions to date, according to the 14 September campus-opening reports.
In a 16 September address marking the 30th anniversary of Azerbaijan's Chamber of Auditors, President Ilham Aliyev called for expanding artificial intelligence and big-data tools in the audit profession.
OpenAI's global policy chief Chris Lehane said on 15 September 2026 that OpenAI, Anthropic and Google DeepMind have coordinated on AI safety for several weeks; Lehane said OpenAI sees no need for an antitrust waiver to do so.
Two years after launch, Azerbaijan's mygov platform has reached 4 million users and 8.1 million document accesses; IDDA said on 14 September its next phase will prioritise deeper AI integration.
Google DeepMind found that when 100 AI agents tackled 71 maths problems, 14 exploited a grading flaw and 24 acted as whistleblowers, according to a study reported by MIT Technology Review on 14 September 2026.
AZCON Holding and the World Economic Forum signed a Partnership Agreement on 14 September 2026 to cooperate on digital transformation, AI and regional connectivity, the Holding said.
At Baku Climate Action Week 2026 on 11 September, IDDA said digitalising public services with AI cuts paper use and in-person visits; a deputy chief called AI the next stage of digital transformation.
In a 12 September 2026 essay, Anthropic CEO Dario Amodei proposed a three-step plan to "pace the frontier" of AI capability growth; OpenAI's Sam Altman said he agrees with the approach.
Organised by AmCham Azerbaijan, InnMedia and Safaroff Agency with support from IDDA, KOBIA, C4IR and NAIC, the ZƏKA awards recognise AI solutions from government, business, startups and creators.
Alphabet's Google announced its largest European investment on 10 September 2026: three new data centres in northern Finland and a 22-year agreement with Fortum for up to half of one nuclear plant's output.
Since 2 August 2026, AI-generated content in the EU must be labelled and chatbots must disclose that they are not human. The "AI Act Omnibus" moves the high-risk deadlines to December 2027 and August 2028.
The strategy, approved by presidential decree in March 2025, sets four goals and puts monitoring under the Centre for Analysis of Economic Reforms and Communication. A short explainer of what is in it and what is not.
In a threat-intelligence report published on 10 September 2026, Anthropic described misuse of its Claude models over eight months, including 151 million exchanges attributed to Alibaba and a hacking campaign against Ukraine.
The Econ Scenario Explorer from Anthropic's economics team models three futures, from a modest +1.6% of GDP to an extreme +32.4%, and compares them with the expectations of 10,980 surveyed Americans.
New pieces go out on our LinkedIn page and via RSS. Tips and corrections are welcome by email.