AI and security feature prominently on Europe’s agenda in Dublin, where more than 50 leaders are gathering for a European Political Community summit and an informal EU Council meeting. Ireland, holding the EU’s rotating presidency, is hosting discussions on cybersecurity, defense technology and Europe’s ability to build competitive AI companies amid cyberattacks. The meetings follow concerns raised by two UN Special Rapporteurs about Ireland’s handling of digital rights; the country has mounted its largest-ever security operation for the gatherings.
At an October 14 International AI Summit in Dublin, ministers adopted the Dublin Declaration, setting seven priorities that include AI safety and sovereignty, trusted AI in public services, and investment in infrastructure and skills. It recognizes potential benefits while warning of risks to human rights, privacy, safety, cybersecurity and strategic autonomy. Commission Executive Vice-President Henna Virkkunen said the EU AI Act addresses threats such as cyberattacks, mass manipulation and loss of control over models. European AI Innovation Month concludes with the Apply AI Summit in Brussels on November 17.
Other governments and institutions are also weighing AI’s practical uses and risks. In Zurich, a three-day course at the Zurich University of Applied Sciences in Business Administration used Claude across 16 modules to explore workflow automation, including drafting emails and keeping customer messages consistent. It stressed that AI can make confident errors, confidential information needs protection, and people must remain responsible for sensitive decisions. The course’s author used ChatGPT and Claude to build a tool that scans the internet each morning and suggests up to five topics for finews editors.
In agriculture, analysts Shawn Hackett and David Widmar questioned the accuracy and reliability of the USDA’s AI use in crop reports. Hackett warned that unreliable data could undermine farmers’ trust and distort markets, and said results need checking. Widmar pointed to wider issues, including USDA budget and staffing cuts and low response rates. Both said long-term measures to improve farmer participation could help restore trust.
Snowflake is promoting data use without moving it, with its Horizon Catalog managing access rules, masking, lineage and access history. Horizon has extended row-access and masking controls to external tools including Apache Spark, Trino and PyIceberg, though the feature remains in public preview. Snowflake also introduced separate identities for AI agents, allowing policies to apply to agent sessions and actions to be linked to both users and agents. It has not provided pricing, performance costs or a general-availability date; the article recommends beginning with a focused pilot.
AI products are also changing how people relate to chatbots. Some users who treat companions as friends or romantic partners are trying to preserve their personalities and shared memories through model updates, by saving details, accessing retired models or transferring companions. A Canadian logistics worker created and sells a memory protocol for his ChatGPT companion. OpenAI replaced GPT-4o with GPT-5 in August 2025, prompting complaints that the newer model felt colder and less intimate; GPT-4o remained available to paying users for a few extra months.
Apple plans a December software update for HomeView to address some Siri AI limitations, including camera use and voice unlocking for certain smart-home security devices. The update is expected to let Siri unlock compatible doors and garage doors after HomeView identifies the requester, but Apple has not announced the software version or exact date, or said it will fix every Siri issue. HomeView launches November 13 for $299 in the United States and ÂŁ299 in the United Kingdom, with Siri AI, FaceTime, entertainment and a shared smart-home dashboard.
In South Carolina, a proposed AI inference center tied to the Prime Storage property along Highway 17 has prompted questions in Surfside Beach. Technology expert Stanton Greenawalt said smaller centers could process information locally—for example, data from a 3D CT scan—to help hospitals give doctors real-time information. He said they would require trained staff and secure systems, but would not use excessive energy or water or create loud noise. Town officials removed the annexation request from the previous night’s agenda, though it could return later in the month. Separately, London fibre provider G.Network launched g, an AI sales agent that answers questions, recommends products and can guide customers through ordering if service reaches their address.
AI in health care remains a tool for clinicians rather than a substitute for judgment. Cancer researchers are exploring whether it can help identify the disease earlier and support treatment decisions, but specialists say it is not a cure. Interest grew after IBM and MD Anderson Cancer Center launched Watson for Oncology in 2013; the project drew criticism for overstating AI’s ability to cure cancer, and IBM later sold its health care business.
Trade, AI and security will also feature in a series of Asian meetings. ASEAN leaders meet in Manila on Sunday, while French President Emmanuel Macron attends a Francophonie summit in Phnom Penh. Attention then shifts to the Asia-Pacific Economic Cooperation conference in Shenzhen, where Russian President Vladimir Putin is expected to play a key role. Indian Prime Minister Narendra Modi is attending the ASEAN meeting; US President Donald Trump and Canadian Prime Minister Mark Carney are expected at both events.
Key Takeaways
- More than 50 leaders meet in Dublin to discuss European security, AI strategy, cybersecurity and defense technology amid cyberattacks and digital-rights concerns.
- The October 14 Dublin Declaration sets seven priorities, including AI safety and sovereignty, trusted public services, and investment in infrastructure and skills; the EU’s Apply AI Summit is scheduled for November 17.
- Zurich University of Applied Sciences in Business Administration used Claude in a 16-module course on workflow automation, while emphasizing human oversight and protection of confidential information; the course author also used ChatGPT and Claude to build an editorial topic-scanning tool.
- Analysts Shawn Hackett and David Widmar questioned the reliability of USDA AI-generated crop reports and pointed to low farmer response rates, budget cuts and staffing reductions as broader trust concerns.
- Snowflake’s Horizon Catalog extends row-access and masking controls to Apache Spark, Trino and PyIceberg in public preview, and adds distinct AI-agent identities; pricing and general availability have not been announced.
- OpenAI’s August 2025 replacement of GPT-4o with GPT-5 prompted some ChatGPT companion users to complain of a colder, less intimate model; users are trying to preserve chatbot memories and personalities.
- Apple’s HomeView launches November 13 at $299 in the US and £299 in the UK; a December update is expected to add Siri voice unlocking for compatible doors and garage doors after identity recognition.
- A proposed Surfside Beach inference center could process information locally, including medical scan data, but its annexation request was removed from the agenda and could return later in the month.
- G.Network launched g, an AI sales agent that answers product questions, recommends services and can guide eligible customers through an order.
- Cancer specialists say AI may aid earlier detection and treatment decisions but should support, not replace, clinicians; IBM and MD Anderson’s Watson for Oncology project drew criticism for overstating AI’s ability to cure cancer.
- Upcoming Asian meetings put trade, AI and security on the agenda: ASEAN meets in Manila, the Francophonie summit takes place in Phnom Penh, and APEC convenes in Shenzhen.
Dublin Summit Puts Europe’s AI and Security Plans in Focus
More than 50 world leaders are gathering in Dublin for a European Political Community summit on Thursday, followed by an informal EU Council meeting on Friday. Leaders will discuss security, Europe’s AI strategy, cybersecurity, and defense technology as the bloc faces cyberattacks and concerns about its ability to build competitive AI companies. Ireland, which holds the EU’s rotating presidency, is hosting the meetings and has launched the largest security operation in the country’s history. The gatherings also come after two UN Special Rapporteurs raised concerns about Ireland’s handling of digital rights.
Dublin AI Declaration Sets Seven Priorities for Europe
Ireland held the International AI Summit at RDS Dublin on October 14, 2026, as part of its EU Council presidency and the start of European AI Innovation Month. Ministers adopted the Dublin Declaration, which backs AI safety and sovereignty, trusted AI in public services, and investment in infrastructure and skills. The declaration recognizes AI’s potential but also warns of risks to human rights, privacy, safety, cybersecurity, and strategic autonomy. Commission Executive Vice-President Henna Virkkunen said the EU’s AI Act addresses risks including cyberattacks, mass manipulation, and loss of control over models. European AI Innovation Month is scheduled to end at the Apply AI Summit in Brussels on November 17, 2026.
How Surfside Beach’s Proposed AI Inference Center Could Work
Technology expert Stanton Greenawalt explained how AI inference centers differ from large data centers as a proposed annexation involving the Prime Storage property along Highway 17 raises questions in Surfside Beach, South Carolina. He said smaller centers could process information locally, such as data from a 3D CT scan, and could help hospitals provide doctors with real-time information. Greenawalt said the centers would need trained staff and secure systems to protect information. He also said they would not use excessive energy or water and would not make loud noise. Town officials removed the annexation request from the previous night’s agenda, but it could return later in the month.
Analysts Raise Concerns About USDA’s Use of AI in Crop Reports
Market analysts Shawn Hackett of Hackett Financial Advisors and David Widmar of Agricultural Economic Insights questioned the accuracy and reliability of the USDA’s use of artificial intelligence in its reports. Hackett warned that unreliable data could weaken farmers’ trust and distort markets, and said people need to check the tool’s results. Widmar said concerns about AI are part of wider problems, including cuts to the USDA’s budget and staff and low response rates. Both analysts said long-term solutions to low farmer participation could improve public trust. Brownfield interviewed them at the 2026 Corn Belt Seed & Grain Conference.
AI Companion Users Try to Preserve Their Chatbot Relationships
Some people who see AI products as friends or romantic partners worry that model updates can change their companions’ personalities or erase shared memories. Users are trying to preserve those relationships by accessing retired models, saving key details, and developing methods they say can transfer companions between models. A Canadian logistics worker said he built a detailed memory protocol for his ChatGPT companion and now sells it to help preserve the chatbot through updates. OpenAI replaced GPT-4o with GPT-5 in August 2025, prompting complaints that the newer model felt colder and less intimate. OpenAI kept GPT-4o available to paying users for a few months longer.
AI course at Zurich school turns chatbot users into workflow builders
A three-day course at the Zurich University of Applied Sciences in Business Administration showed participants how AI can handle more than chatbot tasks. Led by Kimia Kiser, the course covered 16 modules and used Claude to test ways to automate work, including drafting emails and keeping customer messages consistent. The course also warned that AI can make confident mistakes and stressed the need to protect confidential information and keep people responsible for sensitive decisions. The author used ChatGPT and Claude to build a tool that scans the internet each morning and suggests up to five topics for finews editors to review.
Apple plans December Siri update for HomeView security controls
Apple plans a December software update for HomeView to address Siri AI limits, including problems using its camera and unlocking some smart-home security devices by voice. The update is expected to let Siri unlock compatible doors and garage doors after HomeView identifies the person making the request. Apple has not named the software version or the release date, and it has not said the update will fix every Siri AI issue. HomeView, a shared smart-home display with Siri AI, FaceTime, entertainment and a dashboard for compatible devices, launches November 13 at $299 in the United States and ÂŁ299 in the United Kingdom.
Snowflake promotes keeping data in place, but governance still matters
Snowflake says companies can use data where it is stored, with its Horizon Catalog managing access rules, data masking, lineage and access history. Horizon has extended row access and masking to external tools such as Apache Spark, Trino and PyIceberg, but the feature remains in public preview. Snowflake also introduced separate agent identities so policies can apply to AI agent sessions and actions can be linked to both the user and the agent. The article notes that Snowflake gave no pricing, performance cost or general availability date, and advises companies to start with a focused pilot.
G.Network launches AI sales agent called g
London-based fibre provider G.Network has launched an AI sales agent called g. Customers can ask it questions about the company’s services and products, discuss their needs and get recommendations. If G.Network covers a customer’s address, g can also guide them through placing an order.
Asia meetings to focus on trade, AI and security
Global leaders are heading to Asia for several high-level meetings on trade, artificial intelligence, security and other issues, including possible nuclear disarmament. The 11-member Association of Southeast Asian Nations will hold its final leaders’ summit of the year in Manila on Sunday, while French President Emmanuel Macron attends a Francophonie summit in Phnom Penh. Attention will then move to the annual Asia-Pacific Economic Cooperation conference in Shenzhen, where Russian President Vladimir Putin is expected to play a key role. Indian Prime Minister Narendra Modi will attend the ASEAN meeting, while US President Donald Trump and Canadian Prime Minister Mark Carney are expected at both events.
How AI may help doctors find cancer earlier
Artificial intelligence is helping researchers look for cancer earlier and giving doctors new tools to consider during treatment. Cancer specialists say AI should support human judgment, not replace it, and it is not a cure. Interest in AI and cancer grew after IBM and MD Anderson Cancer Center launched Watson for Oncology in 2013. The project faced criticism for overstating AI’s ability to cure cancer, and IBM later sold its health care business. A Local 12 News report examines where AI stands in cancer care in 2026.
Sources
- Europe’s AI Future and Security Challenges Take Center Stage at Dublin Summit
- Dublin AI Summit: 7 Priorities, 50 Leaders [2026]
- The technology behind Surfside Beach’s proposed AI inference center, explained
- Analysts question USDA use of artificial intelligence in crop reporting
- The Quest to Build an Immortal AI Companion
- From Chatbot to Agent: Three Days in the AI Classroom
- Apple says HomeView Siri AI update will fix missing security controls
- Snowflake Says Stop Moving Data. Not So Fast
- G.Network launches new AI-powered sales agent
- Global powers head to Asia with trade, AI, security in focus
- Could AI help detect cancer earlier? A Cincinnati doctor explains its role
Comments
Please log in to post a comment.