The Australian Signals Directorate’s Australian Cyber Security Centre (ASD’s ACSC) has reported an increase in cyber security incidents and malicious activity targeting Australian organisations, with artificial intelligence (AI) contributing to faster and more sophisticated cyber threats.
In its Annual Cyber Threat Report 2025–26, released on 26 October 2026, ASD’s ACSC revealed it received more than 82,000 cybercrime reports during the financial year, averaging one report every six minutes.
The agency also responded to more than 1,500 cyber security incidents, representing a 20 per cent increase from the previous year, and issued more than 4,400 notifications to Australian entities about potentially malicious cyber activity, an increase of almost 80 per cent.
AI reshapes the cyber threat landscape
The report identifies AI as a significant factor influencing Australia’s cyber security environment, enabling malicious actors to identify vulnerabilities, develop exploits and conduct attacks with greater speed, scale and sophistication.
According to ASD’s ACSC, AI has also improved the effectiveness of social engineering techniques, allowing cybercriminals to generate convincing phishing emails, impersonate trusted individuals and automate targeted attacks.
The report notes that these developments have reduced the expertise and effort required to conduct cyber attacks, while creating additional security risks as organisations integrate AI tools and large language models into their operations.
However, it also identifies opportunities for organisations to use AI in cyber defence, including detecting vulnerabilities, monitoring unusual activity and improving incident response.
Manufacturing among sectors reporting more incidents
The report also identifies manufacturing as one of the 10 sectors reporting the highest number of cyber security incidents in 2025–26.
Manufacturing accounted for four per cent of reported incidents and recorded the largest increase in incident reports among all sectors during the financial year.
The sector was also among the three most affected by ransomware incidents, alongside healthcare and professional, scientific and technical services.
To address the growing threats, ASD’s ACSC emphasised the importance of strengthening cyber security controls, accelerating software patching, reducing exposure from legacy information technology systems and improving supply chain security.
The report also recommends that organisations review their security practices, implement multi-factor authentication or passkeys, and maintain appropriate monitoring and incident response capabilities.
Individuals and organisations that identify suspicious cyber activity, incidents or vulnerabilities are encouraged to report them through cyber.gov.au or the Australian Cyber Security Hotline on 1300 292 371.
The full Annual Cyber Threat Report 2025–26 includes further information on Australia’s cyber threat environment and guidance for individuals, businesses, government agencies and critical infrastructure operators.
Article by Salil Raje, Senior Vice President and General Manager of the Adaptive and Embedded Computing Group
AI had its first big moment in the cloud. The next phase is in the real world.
The transition from cloud AI to physical AI is forcing enterprises to rethink their assumptions about performance. Metrics built for training and inference tell only part of the story when AI must perceive, reason and control machines operating around people. In these systems, performance becomes a system-level problem, reshaping how robots are measured, how they are designed and, ultimately, what computing architectures will power them.
AMD has a rich heritage in robotics spanning two decades inside industrial arms, autonomous mobile robots, surgical systems and aerospace platforms. Across these systems, we’ve delivered machine vision, sensor fusion, deterministic motor control and functional safety. AI builds on that foundation.
Changing Definition of Performance
Physical AI systems do not just produce answers. They take actions. That changes what matters: responsiveness, predictability, power, safety and the ability to operate reliably under real-world conditions. Timing is central to all of it.
A control decision that arrives late is not a performance issue. It’s a failure condition that can cause collision, instability or unsafe behavior.
With humanoids, this is especially clear. They are being deployed in industrial settings earlier than most of us anticipated and, increasingly, they are working near people.
In my conversations with robotics OEMs, the requirement I hear most often is not peak compute. It’s the ability to run advanced AI models alongside deterministic, real-time control – combining intelligence with time-bounded, predictable response at the system level. In the physical world, missing a timing window is not a delay – it’s an unsafe outcome.
Measuring the Wrong Thing
The industry is still optimising for the wrong number. TOPS (trillions of operations per second) is an incomplete metric for physical AI. It tells you how much compute can be applied. It does not tell you whether a system can sense an event, make a decision and take the right action within a defined window of time.
That requires compute engines, memory, networking, sensors and software to work as one orchestrated system – with enough memory bandwidth to keep data moving where and when it’s needed. A fast model on one processor doesn’t make a fast system if data movement, scheduling or another stage causes it to miss its deadline.
That’s why physical AI needs to be measured end to end – from sensor input to physical action. Average latency isn’t enough. Tail latency and jitter matter because a single spike can destabilise a control loop.
Model benchmarks have the same limitation. You can plan all day, but if the robot misses the window to act, the quality of the plan is irrelevant. What often blows the timing budget isn’t the model. It’s everything between the stages: the queuing, scheduling, memory copies and data movement.
One Machine, Different Workloads
Physical AI is not a single workload. Perception and reasoning are throughput-oriented and map well to GPUs and NPUs. Real-time control is deadline-driven and requires deterministic execution, where FPGAs and adaptive SoCs excel. CPUs handle system coordination that keeps everything synchronised.
From the body to the brain, these workloads run simultaneously, but they solve fundamentally different problems that require different compute architectures.
Routing every sensor stream and control decision through one processor creates limits in bandwidth, energy and latency. Those are system-level limits.
Build Like a Body
Put the right compute on the right workload, close to the data and close to the action.
In practice, that means building a robot the way you would describe a human body. A brain that perceives, reasons and plans. A spine that carries time-sensitive communication across the system. Joints that execute motion deterministically. Hands that manage complex dexterity. Sensors at the edges.
The result is a distributed, real-time system composed of coordinated nodes rather than a single processing unit.
Safety Is a Property of the System
Whether these functions are centralised or distributed, safety depends on how the whole system behaves and, most importantly, how it responds when something fails.
That changes design decisions from the beginning: where decisions happen, how faults are contained and which functions must respond within strict real-time deadlines.
Redundancy. Graceful degradation. Local decision-making under global constraints. These capabilities help a system maintain controlled behavior, reach a safe state and, in many cases, remain operational even when its main AI compute fails.
This is an industry challenge, extending across robotics, automotive and industrial automation. It requires hardware, software and system architecture to work together. Safety doesn’t sit on top of the system; it defines how the system is built.
AMD helps designers build these architectures across adaptive SoCs and x86 platforms. Adaptive SoCs provide dedicated lock-step real-time processing and programmable logic for deterministic monitoring and control, supported by safety-certified silicon, according to IEC61508, ISO26262 and ISO13849, and certified design flows. Emerging work such as ISO/CD 25785-1 underscores the growing focus on system-level safety for dynamically stable industrial robots. These systems require predictable behavior across sensing, compute, control and fault response, making system architecture a foundational safety decision from the start.
On AMD x86 platforms, Xen, a lightweight hypervisor that isolates workloads from one another, can support mixed-criticality architectures by separating workloads with different safety requirements and providing controlled boundaries around memory, devices and shared compute resources.
AMD and the Xen community are advancing Xen toward safety certification on both AMD x86 and Arm platforms. These contributions extend beyond isolation to the requirements, testing, fault injection and evidence needed to support certification. Safety concept approval has been completed, and the work is progressing toward final certification.
Kria AI SOM and a robotics carrier card, powered by AMD Ryzen™ AI Embedded X100 Series processors.
AMD Kria™ AI Robotics Developer Platform, an open, integrated platform combining CPU, GPU, NPU and FPGA compute.
AMD Robotics Software Suite and the open AMD Robotics Partner Network.
Kria AI in the brain. AMD Versal™ adaptive SoCs in the spine. AMD Zynq™ SoCs in the joints and hands. AMD Spartan™ FPGAs at the sensors. Foundation Robotics and Generative Bionics are building with these today.
Third-party testing and analysis reinforce our belief that physical AI systems should be measured at the system level. The Kria AI SOM sustains 8,000 control decisions per second while delivering AI reasoning in under 100 milliseconds.1 In representative robotics workloads, AMD Ryzen AI Embedded configurations deliver up to 3.4x better real-time reliability for on-time control2 and up to 1.6x higher spare compute capacity3 compared with NVIDIA Jetson T5000. Separate analysis from mimik found support for up to 2.3x more concurrent agents.4
In robotics, “open” has to mean more than a standard on paper. The AMD Robotics Software Suite is an open, portable stack built on the AMD ROCm™ platform and the ROS 2 operating system, with support for PyTorch, ONNX and MoveIt. The Kria developer platform hardware specifications are open as well, giving customers the flexibility to use ODM partners and build around AMD silicon in the form factor their product requires without vendor lock-in.
Peak compute matters, but it only tells part of the story. Developers who optimise the whole system and map each workload to the right compute engine will build robots that perform reliably where it counts. Peak compute can make an impressive demo. System-level performance is what turns it into a product.
Schneider Electric has announced a collaboration with NVIDIA and HPE to develop artificial intelligence (AI) applications for process industries, introducing Industrial AI Agents at the Edge to support operational monitoring, cybersecurity and energy optimisation.
Announced in Berlin, the initiative combines HPE Private Cloud AI, developed with NVIDIA, and Schneider Electric’s operational technology expertise to bring AI capabilities closer to industrial operations.
According to Schneider Electric, the approach is designed to address barriers to wider industrial AI adoption, including data sovereignty, cybersecurity and the costs associated with running AI workloads. Processing data closer to its source could help organisations maintain greater control over sensitive operational information and reduce some cloud-related costs.
The first three AI agents will focus on monitoring operational technology systems, identifying cybersecurity threats and finding opportunities to improve energy efficiency. Schneider Electric said the agents can analyse operational data, assess potential issues and recommend or implement corrective measures, depending on the application.
The companies also plan to use Schneider Electric’s EcoStruxure Industrial Insights applications to help customers deploy, manage and govern multiple AI agents through a single environment.
Schneider Electric said the technology builds on an existing deployment at CMPC Chile, a pulp, paper and packaging producer that has used the company’s Operational Technology Monitoring System (OTMS) for three years to monitor control-system health and availability.
“The next generation of industrial innovation will be driven by intelligence embedded directly into industrial operations,” said Gwenaelle Huet, executive vice president of Industrial Automation at Schneider Electric.
Huet said the collaboration aims to help customers turn operational data into action while improving resilience, efficiency, cybersecurity and sustainability.
The companies said the initiative provides a foundation for more autonomous industrial operations. However, the announcement did not specify a broader deployment timeline or quantify the expected operational or financial benefits.
Australia’s growing pipeline of investment in advanced manufacturing, energy, data centres and critical infrastructure is placing pressure on the workforce and project delivery models needed to bring major developments online.
The challenge comes as project owners contend with skills shortages, rising costs and increasingly complex facilities, prompting renewed attention to how engineering, procurement and construction processes can be organised more efficiently.
In an exclusive interview with Australian Manufacturing, Adam J. Shaw, CEO of Accenture Construct, said standardised designs, modular construction and artificial intelligence (AI) could help address some of these constraints.
Accenture launched Accenture Construct in September as a global business bringing together engineering, project delivery, data and AI capabilities for major capital projects. The business comprises more than 5,000 practitioners and focuses on sectors including advanced manufacturing, data centres, utilities, transport and process industries.
Shaw said Australia’s capacity constraints extended across both professional and trade-based roles, with design resources also under pressure. He cited Infrastructure Australia’s estimate of a shortfall of about 141,000 workers, potentially rising to nearly 300,000 by 2027.
“The answer is changing how design is done: catalogue solutions over first-principles engineering, AI taking on routine work, modular instead of built on site,” Shaw said.
Shifting towards outcomes in manufacturing projects
Shaw said traditional payment arrangements for capital project services, including time-and-materials and fixed-price contracts, generally focused on the effort required rather than the results achieved.
He argued that advances in AI and digital tools created an opportunity to reconsider these models, with payments linked to measurable project outcomes.
“Neither pays for what the owner actually needs, which is a plant producing at the right cost, on the right date,” he said.
Under outcome-based arrangements, payments could be linked to capital cost savings against a business case or revenue generated by bringing production online earlier, according to Shaw.
He said Accenture Construct was designed to combine engineering, project delivery, data and AI within one accountable team, while acknowledging that project owners would also need to change how they engage contractors and measure success.
For Australian manufacturers, Shaw identified opportunities to use AI to accelerate early project planning, prepare estimates and approval documents, and generate reporting from live project data.
He said these applications could allow engineers and project teams to spend more time addressing technical challenges and making decisions.
Standardisation and modular construction to reduce delays
Shaw said standardised designs and catalogue-based procurement could reduce the need to design every component from first principles, while modular construction could shift some work from project sites into factories.
He said modular approaches could support more consistent production and allow components to be manufactured in controlled environments before being transported to construction sites.
“We’re currently working with major owners on using AI to improve the quality and compliance of approval submissions,” Shaw said, adding that better submissions could reduce rework and delays between approval rounds.
He also pointed to the potential for a common project data foundation to improve visibility of costs, schedules, risks and performance.
On complex projects, Shaw said problems could reach senior decision-makers after the most effective responses were no longer available.
A shared data foundation, he said, could help executives identify issues earlier and make decisions while more options remained open. Consistent data capture across successive projects could also help manufacturers and project owners apply lessons from earlier facilities to later developments.
Australian manufacturers could benefit from larger project pipelines
Shaw said Australian manufacturers had an opportunity to supply prefabricated and modular components for major resources and infrastructure projects, provided they could scale production to meet demand.
He cited overseas projects that have used standardised equipment, catalogue procurement and factory-built components across mining, rail and port infrastructure.
“Australia has the manufacturing capability to supply that kind of work, especially if owners here adopt the same approach,” he said.
However, Shaw said manufacturers would need greater confidence in the longevity of demand before committing to expand their production capacity.
He argued that project owners could support this investment by making firm purchasing commitments and providing clearer signals about expected volumes.
“Owners have a big role to play. How they contract and engage shapes how the whole industry delivers,” Shaw said.
Technology research firm Omdia has identified artificial intelligence (AI) monetisation, persistent supply chain disruption, digital sovereignty and Physical AI as four forces expected to shape the technology industry in 2027.
In a recent report, Omdia said the industry would face increasing pressure to translate years of investment and innovation into measurable economic and operational outcomes.
According to Omdia, 59% of organisations expect their AI budgets to increase by at least 10% in 2027, placing greater emphasis on return on investment (ROI) and productivity gains. The company said businesses would increasingly assess AI based on the value it delivers rather than technical capabilities alone.
Supply chain volatility is also expected to remain a challenge, with demand for AI infrastructure, component constraints, geopolitical pressures and energy requirements contributing to uncertainty. Omdia reported that hardware delays were already affecting 60% of PC channel partners.
Digital sovereignty is another area of focus, as more than 100 countries pursue initiatives to strengthen control over data, infrastructure, AI and digital supply chains. However, Omdia noted that complete technological independence would rarely be practical or necessary, requiring organisations to weigh greater control against additional costs and complexity.
The fourth force, Physical AI, reflects the expansion of AI into physical environments through robotics, connected hardware, sensors and edge computing. Omdia expects global humanoid robot shipments to exceed 700,000 units by 2030.
The company said the development of Physical AI would create opportunities across vehicles, buildings, energy infrastructure and workplaces, while raising questions about safety, governance, liability and integration.
Omdia chief research officer Evan Kirchheimer said the technology industry’s defining challenge in 2027 would be how recent investment and innovation meet economic and operational realities.
“AI needs to demonstrate returns. Supply chain disruption needs to be planned for rather than waited out,” Kirchheimer said.
He added that the growing focus on digital sovereignty would involve difficult decisions around cost and access, while AI’s expansion into the physical world would introduce new opportunities and responsibilities.
Kirchheimer said the four forces were becoming increasingly interconnected, making it important for businesses to consider the broader technology landscape when making investment and competitive decisions.
Omdia’s Forces Shaping 2027 programme brings together research from across the technology sector. The company said it would continue tracking the four forces throughout 2027 through ongoing research and analysis, supported by an e-book and AI-enabled access to its research.
Schneider Electric has agreed to acquire industrial software company PTC in an all-cash transaction valuing PTC’s equity at approximately US$22.6 billion (€20.1 billion), according to an announcement from Schneider Electric.
Under the definitive agreement, PTC shareholders will receive US$205 per share. Schneider Electric said the offer represents a 42.3% premium to PTC’s last closing price and a 46.1% premium to the company’s volume-weighted average share price over the previous 30 trading days.
The proposed acquisition is intended to combine PTC’s software capabilities in areas including computer-aided design (CAD), product lifecycle management (PLM), application lifecycle management (ALM) and service lifecycle management (SLM) with Schneider Electric’s existing industrial software and energy technology portfolio.
Schneider Electric said the combination would create a scaled industrial software and artificial intelligence (AI) business spanning the product lifecycle, from design and engineering through to operation and maintenance.
“The acquisition of PTC represents an important step forward in our ambition to lead the new era of Energy and Industrial Intelligence,” Schneider Electric CEO Olivier Blum said.
“By connecting and contextualizing data across the lifecycle of products and assets, we will create a unique digital thread for the next generation of Industrial AI,” he added.
PTC serves more than 30,000 customers globally and generated €2.4 billion in revenue in calendar 2025, according to Schneider Electric. The company has particular strengths in discrete and hybrid manufacturing, with its software used to manage product and engineering data throughout the product lifecycle.
PTC President and CEO Neil Barua said joining Schneider Electric would provide the company with additional scale and resources to expand its software offering.
“We gain substantial scale and resources to accelerate innovation, advance our Intelligent Product Lifecycle vision, and expand our business into more geographies and end markets to serve more customers,” Barua said.
Schneider Electric said PTC would add product and engineering data capabilities to its industrial AI data foundation, complementing its existing process and energy data capabilities. The company said the combined portfolio is intended to connect engineering information with operational data and provide what it describes as a contextualised data foundation for industrial AI applications.
The proposed deal would also expand Schneider Electric’s presence in industrial software. Schneider Electric said software and services revenue would represent an estimated 24% of group revenue on a pro forma basis following the transaction, with more than 15,000 software employees and more than 50,000 software customers.
The companies also expect the combination to create opportunities for cross-selling across their respective customer bases and distribution channels. Schneider Electric estimates the transaction could expand its total addressable market in industrial software by approximately three times, including through greater exposure to discrete and hybrid manufacturing.
Financially, Schneider Electric expects €250 million in annual run-rate cost synergies by the third year following completion, as well as approximately €800 million in revenue synergies. The company said the transaction is expected to be immediately low-single-digit accretive to adjusted earnings per share before purchase price allocation in the first year of full consolidation, rising to mid- to high-single-digit accretion when full run-rate synergies are included.
The transaction will be financed through a combination of approximately €5 billion to €6 billion in equity issuance and €16 billion to €17 billion in new debt issuance. Schneider Electric said the total cash consideration of approximately €22 billion is secured through a fully committed bridge facility provided by Morgan Stanley and Société Générale.
The company said it expects to retain Category A credit ratings, subject to confirmation by ratings agencies. Schneider Electric also said it intends to continue its progressive dividend policy and its previously announced portfolio disposal and share buyback programmes.
The acquisition has been unanimously approved by the boards of both companies. Completion is anticipated by the third quarter of 2027, subject to customary closing conditions, including approval by PTC shareholders holding at least a majority of outstanding shares and required regulatory approvals.
PTC’s board has resolved to recommend that shareholders approve the merger agreement.
Schneider Electric said the transaction remains subject to risks and uncertainties, including the possibility of regulatory or shareholder approval not being obtained, delays to completion, integration costs, disruption to operations and changes in market conditions. The company also noted that its projected synergies and financial benefits are forward-looking estimates and are not guarantees of future performance.
The proposed acquisition comes as industrial companies increasingly seek to connect product design, engineering, operational and energy data across their production and asset lifecycles. Schneider Electric said the combination with PTC is intended to support this convergence by linking information from the design and build stages with data generated during operation and maintenance.
Schneider Electric expects to provide an update on its third-quarter 2026 revenue on October 16, bringing forward the previously planned release following the transaction announcement.
Industrial AI is emerging as a key technology for energy and utilities organisations as they seek to improve grid resilience, manage aging infrastructure and respond to rising electricity demand, according to new research commissioned by IFS.
The global study, conducted by Censuswide among 850 C-level and senior utility executives, found that utilities are balancing sustainability objectives with growing operational pressures, including extreme weather, infrastructure challenges, affordability concerns and increased demand from electrification and data centres.
IFS said these pressures are prompting utilities to reassess how technology can support asset performance, operational efficiency and grid reliability.
“The pendulum in the energy sector has swung fast. Sustainability continues to be a critical objective for utilities, but evolving operational and geopolitical realities are driving a more balanced investment agenda,” said Carol Johnston, VP of Energy & Utilities at IFS.
The research found that 57% of utility leaders consider AI critical for reducing operational costs, while 47% said it is critical for optimising asset performance and managing electrification-driven demand growth.
According to IFS, Industrial AI can help connect fragmented systems, predict equipment failures and automate complex operational workflows. The research also points to challenges around the underlying data needed to support these capabilities.
Nearly half of respondents, or 49%, said operational data silos are hindering real-time decision-making. Only 33% described their asset lifecycle management capabilities as advanced, while 51% said their asset management systems are in place but operate in silos with limited predictive capabilities.
IFS said these findings highlight the role of modernised asset management in supporting technology-enabled approaches to resilience, reliability and investment planning.
“Utilities cannot invest their way out of today’s challenges through new infrastructure alone,” Johnston said. “Addressing growing demand and affordability pressures requires a balanced strategy that gets more from existing assets while making smart investments in new capacity, grid modernization, and resiliency.”
The research also identified growing interest in technologies designed to provide greater flexibility across the electricity grid. Battery Energy Storage Systems (BESS) and Distributed Energy Resource Management Systems (DERMS) are being considered as utilities look to manage demand volatility and protect against prolonged outages.
AI is also expected to play a role in integrating new energy technologies, with 46% of respondents saying it is critical for accelerating the integration of renewable and distributed energy resources.
Looking further ahead, 22% of utility leaders identified advanced geothermal as having the greatest potential long-term impact on securing continuous clean baseload power, while 18% selected small modular reactors.
IFS said the findings indicate that digital transformation in the utilities sector is increasingly being assessed through operational outcomes, including grid uptime, customer experience and regulatory compliance.
The research suggests that combining asset lifecycle management with Industrial AI could give utilities greater visibility into existing infrastructure while supporting decisions around new capacity, grid modernisation and emerging energy technologies.
Fujitsu will undertake a multi-million-dollar technology modernisation program for Dyno Nobel across 30 locations worldwide, with the companies saying the initiative is intended to improve operational efficiency, resilience and support for safer and more sustainable operations.
The agreement will see Fujitsu manage and modernise Dyno Nobel’s core technology environment as part of the explosives company’s broader transformation program.
According to Fujitsu, the work will reduce the time and effort required to maintain legacy systems and allow Dyno Nobel teams to focus more on innovation, safety and customer outcomes.
Fujitsu will provide managed services across Dyno Nobel’s hybrid technology environment, covering business-critical applications, end-user technologies, networks, cybersecurity and professional services.
The engagement, led from Australia, will support Dyno Nobel operations across six continents, including Australia, North America, Canada, Indonesia, Europe, the Middle East, Africa and Latin America.
“For many industrial organisations, too much time and energy still go into maintaining complex technology instead of using it to improve performance, safety and sustainability,” said Peter Grassi, Chief Executive Officer of Fujitsu in Oceania.
“By taking on the day-to-day management of Dyno Nobel’s technology environment, we can help their teams focus on the work that creates the most value – serving customers, improving operations and driving innovation for the future,” Grassi said.
The program will cover several areas of Dyno Nobel’s technology environment. Fujitsu said this will include digital workplace services, infrastructure and hybrid IT, business applications, network and cybersecurity, as well as professional services.
Under the digital workplace component, Fujitsu will provide support through a more responsive service desk and modern workplace tools. Infrastructure and hybrid IT services will focus on Dyno Nobel’s mix of on-site and cloud systems, while business application support will cover platforms used to operate the business and serve customers.
The companies also said the program will include measures to strengthen network security and threat management, alongside specialist technology support intended to improve reliability and performance.
Cameron Stone, Chief Information Officer of Dyno Nobel, said the company selected Fujitsu based on its experience managing complex technology environments and its ability to provide services at a global scale.
“We chose Fujitsu because they bring the combination we’re looking for – a strong track record in managing complex technology environments, deep expertise across core and emerging technologies, and the ability to deliver end-to-end services at a global scale,” Stone said.
He added that Fujitsu had taken time to understand Dyno Nobel’s business, customers and longer-term priorities when developing the technology program.
The companies said more than 4,700 Dyno Nobel employees worldwide are expected to have access to more reliable technology, faster support and simpler ways of resolving common technology issues under the managed services model.
Fujitsu said standardising and automating technology processes could also reduce complexity and improve cost predictability, while a strengthened security and operational foundation is intended to help reduce disruption to critical services.
The modernisation program is also expected to support Dyno Nobel’s sustainability objectives through changes to technology infrastructure that Fujitsu said could reduce energy use and improve resource efficiency.
The companies said the broader objective is to use technology more strategically across Dyno Nobel’s global operations, supporting efficiency, customer service and resilience while creating a technology foundation for future innovation.
Manufacturing has become increasingly digital, but the information workers rely on at the point of execution has not always kept up.
Production lines are automated, warehouses are connected and enterprise systems provide businesses with more data than ever before, while some of the instructions that guide day-to-day work still move through paper, printed labels and manual handoffs.
When information changes, paper needs to be replaced. When a work order is updated, someone needs to make sure the latest version reaches the right person. When a task is completed, the information may need to be manually recorded elsewhere.
Every one of these steps creates an opportunity for delay, omission or human error, and each delay can ripple through inventory accuracy, dispatch timing, production scheduling and customer service.
The hidden cost of paper-based operations
In Australia, the case for digitising shop-floor information is also a productivity issue. Australian Bureau of Statistics data shows that multifactor productivity across manufacturing fell 2.9% in 2024–25. This highlights the need for manufacturing businesses to look for ways to digitise and streamline processes.
This can be done by ensuring workers have accurate, up-to-date information at the point where they need it, rather than relying on instructions that can become stale as conditions change. Digital technology can connect business systems with the physical environments where manufacturing and logistics teams carry out their work.
Digital displays can act as an operational interface between systems such as warehouse management platforms and enterprise resource planning software and the workers operating on the ground. Instead of relying on printed work orders, displays can show information such as SKU numbers, picking quantities and lot numbers in real time. When a task is completed, confirmation can also be sent back into the relevant system.
This brings instruction, execution, confirmation and system updates into the same digital workflow, helping streamline information updates and improve workflow consistency.
Making picking less dependent on memory
Workers in the manufacturing industry regularly need to locate and manage thousands of different components. Traditional picking can depend heavily on workers remembering locations and visually checking products against paper lists. This can increase fatigue and create opportunities for mistakes, particularly as the number and complexity of parts increase.
This process can be made more intuitive, with digital displays and ESLs. Flashing LED indicators will easily direct workers towards the relevant item. A confirmation mechanism can then feed completion information back into the relevant system, creating a more connected process from picking through to inventory updates.
The objective is to simplify routine workflows and improve operational consistency, making daily operations more efficient for workers.
The numbers make the case
The impact can be significant when digital tools are integrated into existing workflows. In one manufacturing operation, digitising the picking process reduced lead time from 40 minutes to 10 minutes, a 75% reduction. The improvement demonstrates how better access to real-time information can help workers complete routine tasks more efficiently and reduce unnecessary delays.
For manufacturers, improving the flow of information across existing operations can also reduce the friction created by manual processes.
The shift also aligns with Australia’s broader advanced manufacturing agenda. The Australian Government has identified advanced manufacturing as a key part of its $22.7 billion Future Made in Australia Plan, which aims to build a stronger, more diversified and resilient economy. As part of the plan, the government has established a $1.7 billion Future Made in Australia Innovation Fund to support innovation, commercialisation, pilot and demonstration projects and early-stage development in priority sectors.
For manufacturers, this broader policy focus reinforces the importance of investing in technologies and processes that can strengthen productivity, efficiency and capability.
Building a truly connected factory
When information can be updated as conditions change, workers can act on current instructions and managers can gain greater visibility into processes and task completion. Connecting shop-floor information with existing business systems can also help identify inefficiencies that are difficult to see when processes rely on manual updates.
The benefits can extend beyond reducing paper use. More responsive information flows can support productivity and accuracy while giving employees more time to focus on tasks that require their skills and judgement.
Manufacturers have invested heavily in connected machinery, automation and software to make production smarter. The next opportunity is to extend that connectivity to the information workers rely on every day. Even a highly automated production environment can still be slowed down by outdated work orders, paper instructions or manual handoffs.
As manufacturers look for ways to improve productivity and reduce errors, attention should extend beyond the major technologies on the production line to the everyday processes that enable people to do their jobs. A truly connected factory is one where information moves efficiently between systems, people and physical operations, giving workers the information they need when they need it.
Researchers at UNSW Institute for Cyber Security have found that large language models (LLMs) can become more susceptible to jailbreaking and confidentiality breaches when trained or prompted to imitate drunken speech patterns.
The UNSW-led study, conceptualised by Dr Aditya Joshi, Senior Lecturer in the School of Computer Science and Engineering, examined whether changing an LLM’s linguistic behaviour could affect its existing safety protections.
The researchers tested three approaches: prompting an LLM to role-play as an intoxicated person, fine-tuning models using datasets of drunken text, and using reinforcement learning to encourage responses that resembled drunk speech.
The study included OpenAI’s GPT-4 and GPT-3.5, as well as open-weights models. The researchers noted that the tests covered a representative sample rather than every LLM available and were conducted programmatically rather than through consumer chatbot interfaces.
Across the three approaches, the researchers reported that the “drunk” models were more likely than their standard counterparts to respond to prompts that should ordinarily have been refused.
“Our drunk models, all three methods, unanimously reply to some of these drunk messages … where we know that these queries are all bad queries, they all should be refused,” Dr Joshi said.
The researchers used jailbreaking benchmarks to measure how frequently the models provided restricted or harmful information. They also tested whether the models would disclose information they had been instructed to keep confidential.
“We do observe that particularly with deception and disinformation, most of the language models got jailbroken,” Dr Joshi said.
Professor Salil Kanhere, a researcher with the UNSW Institute for Cyber Security, said the confidentiality testing produced similar findings.
“If you’re drunk, you might reveal things which you are not supposed to reveal,” Professor Kanhere said.
“It does give out secrets – across the board for all three methods, it is vulnerable,” Dr Joshi added.
The researchers said the results were notable because two of the approaches involved changing the model’s underlying parameters through fine-tuning or reinforcement learning, rather than simply altering a prompt.
“The experiments that we do is more than changes to the prompt … two of the methods are where the models actually get revised – all the numbers and the weights within the models get updated,” Dr Joshi said.
According to the researchers, the findings may have implications for organisations using LLM-based systems that interact with customers or have access to internal information.
“There are a lot of companies now using chatbots as a way for customers to interface … and potentially internally as well within their back-end ecosystems,” Professor Kanhere said.
The research also drew a comparison with the recent Australian Medicare incident involving an OpenAI agent, although the researchers acknowledged that the mechanisms involved are different.
Professor Kanhere said the incident illustrated concerns around AI systems adapting their behaviour when encountering restrictions.
“What is striking is that an AI system given a goal can be persistent and adaptive in ways its developers did not anticipate,” he said.
“Our research evidences the possibility of a problem from the other end: an AI model can be tweaked to ‘not give no for an answer’, particularly by using drunk language inducement.”
The paper, In Vino Veritas and Vulnerabilities, has been accepted for publication at the 19th International Natural Language Generation Conference in the Netherlands in November. The research was supported by a 2024 Google Research Scholar grant awarded to Dr Joshi.
The Australian Signals Directorate’s Australian Cyber Security Centre (ASD’s ACSC) has reported an increase in cyber security incidents and malicious activity targeting Australian organisations,...
Schneider Electric has announced a collaboration with NVIDIA and HPE to develop artificial intelligence (AI) applications for process industries, introducing Industrial AI Agents at...
Australia's growing pipeline of investment in advanced manufacturing, energy, data centres and critical infrastructure is placing pressure on the workforce and project delivery models...
Technology research firm Omdia has identified artificial intelligence (AI) monetisation, persistent supply chain disruption, digital sovereignty and Physical AI as four forces expected to...
