Experts are warning that artificial intelligence could erode specialized technical skills in industries like packaging. Retired packaging leader Jay Gouliard, who spent decades at companies like Coca-Cola, Unilever, and Mondelēz, says AI might reduce hands-on expertise because no two production lines are the same. He believes fully automated packaging design remains years away, possibly around 2050, and argues AI should assist rather than replace human workers.
Meanwhile, AI is also reshaping how people interact with technology on a personal level. A West Chester University student writer explores how users assign gender to AI bots like ChatGPT, often referring to them as he or him despite the technology having no gender. The article says this reflects deeper societal patriarchy and calls for reflection on gender stereotypes as AI becomes more common.
In Japan, autumn 2026 saw multiple data leaks tied to AI-assisted hacking. Trend Micro researchers studied three attack cases and found AI played different roles in each. One involved CyberStrikeAI, an open-source platform with 44 AI agents handling reconnaissance and data theft. Another case featured a human attacker using a large language model to plan strikes on AI infrastructure. Japan's National Cybersecurity Office declared cyberspace a state of emergency and warned businesses.
On the infrastructure side, Ethernet technology is proving critical to scaling AI training systems. Technologies like RoCEv2, Priority Flow Control, and Explicit Congestion Notification help Ethernet manage synchronized GPU traffic more efficiently. New dynamic load balancing methods allow packets to use multiple paths at once, reducing congestion as AI clusters continue to grow.
In Southeast Asia, nuclear and radiological facilities are becoming more digital, but governance rules have not kept pace. Julius Cesar Trajano of the S. Rajaratnam School of International Studies argues that cyber security must be part of nuclear security. The region lacks a clear nuclear-specific cyber regime, and risks include ransomware, remote vendor access, and weak coordination between agencies. Joint exercises and stronger institutional coordination are being called for.
Connecticut is pushing to build AI workforce capacity. Gov. Ned Lamont and state officials highlighted the Tech Talent Accelerator, a four-year-old program connecting higher education institutions with businesses. A software engineer at CGI shared how an early internship led to a full-time career and now mentorship opportunities. Lamont expressed optimism that AI knowledge-based work will drive future economic growth in the state.
Arize rebuilt its AI agent Alex by focusing on better harness code rather than relying on improved models. Fuad Ali said the team started in late 2023 with a prototype that often hallucinated. They settled on Alex 2.0 in February 2026 and One Alex in June, improving planning, memory, and task handling through better engineering design. The current default model is Sonar 4.5, and customers can choose their own provider.
AI is also speeding up product development, making technology partnerships more important than ever. Customers need trusted advisors to navigate fast-changing tools, and traditional channel programs with quarterly training no longer keep up. Companies like Censys are using hands-on methods such as sandbox access and gamified learning to help partners stay current. Partners are now key sources of innovation and customer insight.
A round-table discussion at COBEM 2025 explored AI opportunities and challenges in mechanical engineering. The study stresses that AI should complement human engineering judgment rather than replace it. Key concerns include energy use, validation, uncertainty handling, and ethical responsibility in safety-critical applications. Researchers call for curriculum reform and integration of physical science knowledge into AI systems.
Key Takeaways
- Japan declared cyberspace a state of emergency after autumn 2026 data leaks linked to AI-assisted hacking, including attacks using a 44-agent platform called CyberStrikeAI.
- Arize rebuilt its AI agent Alex using better harness code design, settling on Alex 2.0 in February 2026 and One Alex in June, with Sonar 4.5 as the default model.
- Connecticut's Tech Talent Accelerator program connects higher education with businesses to build AI workforce skills, with Gov. Ned Lamont backing AI knowledge-based economic growth.
- Retired packaging leader Jay Gouliard warns AI could erode industry-specific skills and says fully automated packaging design is likely decades away, around 2050.
- Ethernet technologies like RoCEv2 and dynamic load balancing are critical for managing synchronized GPU traffic in large AI training clusters.
- Southeast Asia lacks a nuclear-specific cyber security regime, with risks including ransomware, remote vendor access, and poor coordination between cyber agencies and nuclear regulators.
- Students at West Chester University observe that users assign gender to AI bots like ChatGPT, reflecting deeper societal gender stereotypes.
- A COBEM 2025 round table calls for AI in mechanical engineering to complement human judgment, with concerns around energy use, validation, and ethics in safety-critical settings.
- Technology partnerships are becoming essential as AI accelerates product development, with companies like Censys using sandbox access and gamified learning to keep partners up to date.
- Trend Micro researchers found AI played different roles in Japan's hacking cases, from autonomous agent platforms to human-led planning using large language models.
AI Could Erase Packaging Skills Experts Warn
Retired packaging leader Jay Gouliard warns that AI might reduce technical skills in the packaging industry. He worked for decades at companies like Coca-Cola, Unilever, and Mondelēz. He says no two production lines are the same, making full digital twins very hard to build. He believes AI should assist humans rather than replace them. He thinks fully automated packaging design is still many years away, possibly around 2050.
Japan Cyberattacks Reveal Different AI Hacking Methods
In autumn 2026, Japan reported multiple data leaks linked to AI-assisted hacking. Trend Micro researchers studied three attack cases and found AI played different roles in each. One case used CyberStrikeAI, an open-source platform with 44 AI agents handling tasks like reconnaissance and data theft. Another case involved a human attacker using a large language model to plan attacks on AI infrastructure. Japan's National Cybersecurity Office warned businesses and declared cyberspace a state of emergency.
AI Era Demands Stronger Technology Partnerships
AI is speeding up product development, making partnerships more important than ever. Customers need trusted advisors to navigate fast-changing technology. Partners help connect different tools and bring real expertise to customer problems. Traditional channel programs with quarterly training no longer keep up with rapid updates. Companies like Censys are using hands-on methods such as sandbox access and gamified learning to help partners stay ready. The article argues partners are now key sources of innovation and customer insight.
AI Governance Gaps Threaten Nuclear Security in Southeast Asia
Southeast Asia's nuclear and radiological facilities are becoming more digital, but rules have not kept pace. Julius Cesar Trajano of the S. Rajaratnam School of International Studies argues that cyber security must be part of nuclear security. The region lacks a clear nuclear-specific cyber security regime. Risks include ransomware, remote vendor access, and weak coordination between cyber agencies and nuclear regulators. The article calls for joint exercises and stronger institutional coordination to protect critical facilities.
Arize Rebuilt Its AI Agent Using Better Code Design
Arize rebuilt its AI agent Alex by focusing on harness code rather than relying on better models. Fuad Ali said the team started in late 2023 with a prototype that often hallucinated. They tried several designs before settling on Alex 2.0 in February 2026 and One Alex in June. The team improved planning, memory, and task handling by building better harness systems. The current default model is Sonar 4.5, and Arize lets customers choose their own provider. The article highlights that strong engineering design can matter more than model upgrades.
Students explore gender bias in AI chatbots
A West Chester University student writer examines how people assign gender to AI bots like ChatGPT. Many users refer to AI as he or him even though the technology has no gender. The author notes that people form personal connections with chatbots and that this bias reflects deeper patriarchy in society. The article calls for reflection on gender stereotypes as AI becomes more common.
Ethernet technology key to scaling AI training systems
AI training requires thousands of GPUs to work together, but packet loss and network delays can slow the entire system. Technologies like RoCEv2, Priority Flow Control and Explicit Congestion Notification help Ethernet handle synchronized GPU traffic more efficiently. New dynamic load balancing methods allow packets to use multiple paths at once to reduce congestion. Ethernet continues to evolve to meet the growing demands of large AI clusters.
Connecticut leaders promote AI workforce training programs
Connecticut Gov. Ned Lamont and state officials highlighted workforce development programs focused on AI skills. The Tech Talent Accelerator is a four-year-old program that connects higher education institutions with businesses. A software engineer at CGI shared how an early internship led to a full-time career and now mentorship opportunities. Lamont expressed optimism that AI knowledge-based work will drive future economic growth in the state.
Study outlines AI opportunities and challenges in mechanical engineering
A round-table discussion at COBEM 2025 explored how AI is reshaping mechanical engineering across modeling, simulation and design. The study stresses that AI should complement human engineering judgment rather than replace it. Key concerns include energy use, validation, uncertainty handling and ethical responsibility in safety-critical applications. Researchers call for curriculum reform and the integration of physical science knowledge into AI systems.
Sources
- AI has a Drawback Packaging Professionals Should Know
- Three Japan-Focused Cyberattacks Show How Hackers Use AI
- Why We Are Rethinking Partnerships in the AI Era
- AI and cyber governance lessons could shore up nuclear security in Southeast Asia
- Arize rebuilt its agent around harness code not model magic
- Uncovering Gender Bias In Artificial Intelligence
- Taming the AI Training Beast: Why Ethernet Will Write the Next Chapter of AI
- CT leaders tout efforts to train workforce in AI
- Artificial Intelligence in Mechanical Engineering: Foundations, Opportunities, Challenges, and Responsibilities
Comments
Please log in to post a comment.