Cofrai, a Madrid-based startup, has raised €2 million in seed funding to expand its digital platform for the fire protection industry.
AI system detects fires before alarms sound, NYU study shows
The NYU Tandon School of Engineering has reported that its Fire Research Group has developed an artificial intelligence system that can detect fires and smoke in real time using existing CCTV cameras.
NFPA survey highlights AI and training in skilled trades
The NFPA 2025 survey highlights AI adoption, digital tools, training demand, certification gaps, and workforce pressures in the skilled trades sector.
Zelim partners with US Navy on AI detection research
Agreement enables joint trials with US Navy
Zelim has announced a Cooperative Research and Development Agreement (CRADA) with the US Navy’s Naval Undersea Warfare Center (NUWC) Division Newport to test its AI detection system, ZOE Shield.
According to Zelim, the agreement allows the company to participate in BlueTIDE 2025, a demonstration event organised by 401 Tech Bridge, NavalX and the Northeast Tech Bridge.
The company explained that the event will conclude with a full-scale in-water trial on 28 August in Narragansett Bay, Rhode Island.
Zelim said it is one of a small number of international companies selected for the 2025 BlueTIDE Prize Challenge following a competitive review process.
The company stated that this year’s mission scenario focuses on protecting subsea infrastructure from threats posed by small crewed and uncrewed autonomous systems.
ZOE Shield trial and operational role
Zelim reported that its role in the scenario is to detect small remotely operated surface vessels operating covertly in sensitive areas.
It said these craft may present a potential threat to subsea infrastructure by serving as a precursor to sabotage.
The company explained that its ZOE Shield system is based on the same AI engine used in its ZOE man overboard system, which is deployed in the cruise and offshore sectors.
It added that ZOE Shield extends these functions by providing situational awareness of small surface craft not using AIS.
The developer outlined that the system introduces automated classification and alerting, helping operators with early detection and response measures.
Features of the AI detection system
According to Zelim, ZOE Shield is designed to function in environments where GPS denial or radar jamming could limit conventional surveillance methods.
It stated that this makes the system relevant for security operations requiring layered detection and resilience against interference.
The company noted that the US Navy is evaluating the system as part of its assessment of intelligent surveillance technologies.
Zelim explained that this evaluation will take place in live operational scenarios during BlueTIDE.
It added that the data will be used to support the Navy’s capability review and potential future development.
CEO comments on defence scenario
Sam Mayall, CEO and co-founder of Zelim, said: “This agreement gives us a unique opportunity to trial ZOE Shield in an operational defence scenario at a US Navy facility, where multiple assets, including USVs will be deployed on the mission.
“This is about closing the gap between detection and decision, especially in domains where the threat is small, fast-moving and unpredictable.
“Our mission is to help operators identify anomalous behaviour earlier, whether the goal is rescue or protection.
“This demo is a vital step toward deploying ZOE for real-world defence scenarios.”
Role of CRADAs in research collaboration
Zelim stated that CRADAs enable non-federal organisations to work directly with US Navy personnel and facilities.
It said this framework allows research partners to collaborate outside of federal acquisition rules.
The company explained that the structure also protects intellectual property and proprietary data.
According to Zelim, this mechanism is being used to facilitate the ZOE Shield trial in Rhode Island.
It added that further information will be released in line with CRADA requirements after the demonstration.
Relevance for fire and safety professionals
AI detection technologies such as ZOE Shield have potential applications in safety-critical fields beyond defence.
The system’s ability to function without reliance on radar or GPS may inform the design of fire and rescue monitoring technologies in environments subject to interference.
The early classification and alerting features under test may also have relevance to operators seeking to improve situational awareness in complex emergency scenarios.
Zelim partners with US Navy on AI detection research: Summary
Zelim has signed a CRADA with the US Navy’s NUWC Division Newport to test its AI detection system ZOE Shield.
The agreement enables Zelim’s participation in BlueTIDE 2025, a demonstration event scheduled to conclude on 28 August in Narragansett Bay, Rhode Island.
Zelim is one of a small number of international companies selected for the 2025 BlueTIDE Prize Challenge.
The company will trial ZOE Shield to detect small surface vessels operating covertly near subsea infrastructure.
The AI engine used in ZOE Shield is also employed in Zelim’s ZOE man overboard system, which is already deployed in maritime industries.
ZOE Shield provides automated classification and alerting functions and is resistant to GPS denial and radar jamming.
The system is being evaluated by the US Navy for its role in layered maritime surveillance and response.
Sam Mayall, CEO and co-founder of Zelim, said the demonstration is an opportunity to test ZOE Shield in a live defence scenario.
CRADAs enable collaboration with the US Navy while protecting intellectual property and removing federal acquisition requirements.
Further details will be made available after the demonstration in line with the CRADA framework.
AI meets civil defence: Abu Dhabi’s new safety pact could redefine protection
Agreement aims to enhance safety in Abu Dhabi
The Abu Dhabi Media Office has reported that the Abu Dhabi Civil Defence Authority has signed a cooperation agreement with the Advanced Technology Research Council to advance safety and civil protection in the emirate.
The authority said the partnership will focus on developing artificial intelligence strategies, training national talent, and integrating smart solutions into civil defence operations.
According to the announcement, the agreement aims to create sustainable systems that can respond effectively to current and future safety challenges.
The agreement was signed by Brigadier General Salem Abdullah bin Barak Al Dhaheri, Director General of Abu Dhabi Civil Defence Authority, and His Excellency Shihab Issa Abu Shihab, Director General of the Advanced Technology Research Council.
The Abu Dhabi Media Office said the initiative aligns with efforts to enhance operational efficiency and effectiveness in civil defence services.
Strategic goals for AI integration in civil defence
Brigadier General Al Dhaheri said: “This partnership marks a significant milestone in advancing prevention and safety tools and technologies.”
He added: “The integration of artificial intelligence into civil defence operations reflects the authority’s commitment to keeping pace with global advancements and leveraging them to serve the community while enhancing readiness to face a wide range of challenges.”
The authority stated that AI integration will be used to support prevention, rapid response, and operational decision-making.
It added that the approach is designed to improve resource allocation and the speed of emergency response.
The Abu Dhabi Media Office said the partnership is part of a broader plan to develop a modernised civil protection ecosystem.
Research council’s role in safety technology development
His Excellency Abu Shihab said: “This collaboration represents a major step towards leveraging research and development to serve Abu Dhabi’s strategic priorities.”
He added: “The agreement strengthens the research ecosystem’s ability to translate advanced technologies into tangible solutions that enhance community safety and protection, reaffirming the council’s role as a key driver of smart innovations in Abu Dhabi.”
The council said it will contribute through applied research projects and technology testing.
It noted that collaboration with operational agencies such as the civil defence authority will accelerate the deployment of tested solutions.
According to the Abu Dhabi Media Office, the council’s remit includes support for the security and defence technology sectors.
Building a smart and adaptive safety ecosystem
The Abu Dhabi Civil Defence Authority said the cooperation will focus on creating a flexible and effective civil protection system.
It added that the system will be capable of adapting to change, minimising risks, and maintaining high operational standards.
The authority stated that investment in knowledge and advanced tools will help address evolving safety threats.
According to the Abu Dhabi Media Office, the aim is to integrate prevention measures with rapid-response capabilities.
It said this will help protect lives and property while maintaining quality of life.
Future projects under the agreement
The Abu Dhabi Media Office reported that the agreement includes joint work on innovative projects and smart safety solutions.
It stated that AI technologies will be applied to community prevention measures, risk analysis, and operational support systems.
The authority said project outcomes will be designed for scalability across Abu Dhabi.
The research council added that project evaluation will focus on measurable improvements in safety performance.
It noted that results will be shared with relevant government bodies to inform policy and planning.
Abu Dhabi civil defence and research council partner to advance safety through AI: Summary
The Abu Dhabi Media Office has reported that the Abu Dhabi Civil Defence Authority and the Advanced Technology Research Council have signed a cooperation agreement to advance safety in Abu Dhabi.
The agreement focuses on artificial intelligence, integrated strategies, and training specialised national talent.
It was signed by Brigadier General Salem Abdullah bin Barak Al Dhaheri and His Excellency Shihab Issa Abu Shihab.
The partnership will work on sustainable safety systems capable of addressing current and future challenges.
Brigadier General Al Dhaheri said the agreement advances prevention and safety technologies.
His Excellency Abu Shihab said it will translate advanced technologies into community safety solutions.
The cooperation includes projects to create adaptable civil protection systems.
AI applications will be used for prevention, risk analysis, and operational decision-making.
The research council will focus on applied research and technology testing.
Results will be shared with government bodies to support planning.
BC Wildfire Service warns of artificial intelligence wildfire images
BC Wildfire Service flags AI wildfire photos
The Canadian Press has reported that the BC Wildfire Service is warning about artificial intelligence-generated images being shared online during wildfire season.
The service said the images can cause misinformation and confusion when people are seeking updates on active blazes.
It explained that two images circulating on social media were confirmed to be generated with artificial intelligence.
According to the service, the images appeared to depict large-scale firefighting operations involving aircraft.
It stated that these images did not accurately reflect the terrain, fire size, or behaviour of the blazes they appeared to show.
Example of Drought Hill fire image
The BC Wildfire Service said one of the images was first posted by a self-described digital creator on Facebook on 31 July.
The post was accompanied by a caption referencing the Drought Hill fire near Peachland.
The following day, the caption was edited to include a disclaimer that the image was artificial intelligence-generated and intended for illustrative purposes only.
The service noted that people scrolling past these images could assume they were real.
It stated that this misperception could spread inaccurate information about current fire events.
Impact of misinformation during wildfire season
The BC Wildfire Service said people often turn to social media for updates when wildfires are active.
It added that whether such images are shared with good intentions or to mislead, misinformation is a problem in an emergency.
The service said inaccurate information can cause uncertainty among the public.
It stated that emergency response operations benefit from accurate and timely updates.
It urged users to verify the origin of images before sharing them.
Evacuation alert lifted in Peachland
The Canadian Press reported that the Regional District of Central Okanagan has lifted an evacuation alert for 118 properties in the Peachland area.
This follows containment efforts that brought the Drought Hill fire under control.
The BC Wildfire Service confirmed that the status change means no properties remain under evacuation alert in the affected zone.
It added that fire crews continue to monitor the site for hot spots.
Authorities stated that weather and on-the-ground conditions will determine future updates.
Official statement on emergency communications
The BC Wildfire Service said it aims to ensure public updates are based on verified information.
It encouraged residents to follow official channels for the most reliable details on active wildfires.
The service explained that visual media plays a powerful role in shaping public understanding of emergencies.
It stated that artificial intelligence-generated content can interfere with that process if it is not clearly labelled.
The organisation said it will continue to identify and correct instances of misinformation during the fire season.
Relevance for fire and safety professionals
For fire and safety professionals, the warning highlights the need to address misinformation in public communications during emergency operations.
It shows how artificial intelligence-generated imagery can influence public perception if it is mistaken for authentic documentation.
It reinforces the importance of verifying sources before sharing operational updates online.
It also indicates that emergency agencies may need to increase their digital literacy monitoring and media verification capabilities during wildfire season.
BC Wildfire Service warns of artificial intelligence wildfire images: Summary
The Canadian Press reported that the BC Wildfire Service has issued a warning about artificial intelligence-generated wildfire images being shared online.
The service said two images confirmed to be AI-generated were circulating on social media.
It stated that the images did not reflect the real terrain, fire size, or fire behaviour of the blazes they appeared to show.
One image referenced the Drought Hill fire near Peachland and was later edited with a disclaimer.
The service said people often rely on social media for wildfire updates and misinformation is problematic.
It stated that misleading images can create public uncertainty during emergencies.
The Regional District of Central Okanagan lifted an evacuation alert for 118 properties near Peachland.
The service urged the public to use official sources for accurate wildfire information.
Motorola Solutions introduces ‘AI nutrition labels’ for public safety technologies
New labels clarify AI usage in Motorola Solutions’ security systems
Motorola Solutions has announced the launch of ‘AI nutrition labels’ to explain how artificial intelligence is used in its safety and security technologies.
According to Motorola Solutions, the new labelling system outlines core components of each product’s AI features, including the type of AI applied, data ownership, human oversight, and the intended use of AI within the product.
The company stated that this initiative is the first of its kind for public safety and enterprise security products.
The labelling approach has been compared to food nutrition labels, providing transparent and concise information for users of AI-enabled technology.
Motorola Solutions said the labels aim to help users better understand how AI is deployed in tools used by emergency services, businesses, and other critical sectors.
Labels aim to explain purpose, control and context
Each AI nutrition label will describe the specific AI model or method in use, including whether it involves machine learning, deep learning, or rule-based systems.
Motorola Solutions explained that the labels also identify who owns the data being processed and what human input or oversight is present in each product.
The initiative is led by the company’s Technology Advisory Committee, which serves as a cross-functional group advising on the ethical and technical aspects of product development.
The company said this committee acts as its “technical conscience” in relation to product ethics, constraints, and broader implications.
According to Motorola Solutions, these labels are intended to provide clarity to end users on how AI functions in each specific context, especially in high-pressure situations where time and information accuracy are critical.
AI technologies to assist, not replace, human judgement
Motorola Solutions stated that the broader AI strategy focuses on “assisted” intelligence, not full automation.
The aim is to provide accurate and actionable information that helps users prioritise tasks during emergencies or complex events.
Mahesh Saptharishi, executive vice president and chief technology officer at Motorola Solutions, said: “It is our unwavering conviction that technology – including AI – is the bedrock for safety and security, and it must be deployed with purpose and transparency to fulfil its promise as a force for good.”
He added: “Nutrition labels help describe AI’s use in protecting neighbourhoods and nations, and we are proud to take a lead role in bringing greater transparency to AI innovation.”
Motorola Solutions links transparency to effective response
The company stated that AI plays a key role in its safety ecosystem by supporting people with information drawn from varied sources.
It said AI can process and relate data faster than individuals, helping responders make informed decisions in high-risk situations.
Saptharishi said: “Safety threats often unfold at a scale, speed and sophistication that can outstrip any one person’s capacity to make sense of the situation.”
He added: “AI can ingest, learn and cross-reference data to provide contextual understanding.
“At Motorola Solutions, we design our AI-enabled technologies to augment human focus, effort and performance when seconds matter most.
“Our AI nutrition labels will bring added clarity to the important role AI is playing in helping to protect people, property and places.”
Initiative aligns with company’s ethical oversight model
Motorola Solutions said the AI nutrition labels are part of a wider commitment to responsible technology development.
The initiative has been guided by the Motorola Solutions Technology Advisory Committee (MTAC).
The company said MTAC advises on the ethics, limitations and implications of AI and other emerging technologies used in its products.
It also stated that MTAC plays a role in reviewing and influencing product design to align with broader safety and privacy standards.
Motorola Solutions introduces ‘AI nutrition labels’ for public safety technologies: Summary
Motorola Solutions has announced a new labelling system to explain the use of AI in its safety and security products.
The company described the system as ‘AI nutrition labels’, modelled after food packaging disclosures.
Each label includes information on the type of AI used, data ownership, human control, and the purpose of the technology.
The initiative is the first of its kind in public safety and enterprise security.
Motorola Solutions said the labels were developed to increase transparency and user understanding.
The system is part of a broader assisted intelligence strategy focused on supporting decision-making.
Motorola Solutions stated that AI should help users interpret complex data rather than replace human judgement.
Mahesh Saptharishi, Motorola Solutions’ CTO, said the labels demonstrate how AI contributes to safety outcomes.
The company said the labels were developed by its Technology Advisory Committee.
This internal group advises on ethical and technical aspects of product development.
Motorola Solutions said the labels will appear across its AI-enabled product range.
Honeywell AI platform launches for US building management
Early adopters trial Honeywell Connected Solutions platform
Honeywell has announced the launch of its AI-based platform, Honeywell Connected Solutions, for managing building systems in the United States.
According to Honeywell, the platform allows users to integrate building software, systems and devices into one interface. The company said this helps reduce labour time, costs and operational disruption.
Honeywell said the platform has already been deployed by Verizon Communications and Vanderbilt University. These early adopters are using the system to identify potential faults, improve energy management and streamline operations.
The solution is built on the Honeywell Forge platform and includes tools for remote monitoring, predictive maintenance and energy tracking. Installation is supported by AI and can be completed within hours.
New system aims to simplify building management
Honeywell stated that the platform provides real-time visibility into key systems across a building or campus.
The company said building managers can use the system to manage all connected Honeywell equipment, improving coordination between systems that typically operate in isolation.
Honeywell also explained that its integrated interface can assist with compliance, monitoring, troubleshooting and incident response.
The AI-assisted setup reduces the time and cost typically associated with installation of new building management systems.
According to the company, the system’s design supports multi-site operators, integrators and facility managers seeking consistent operational insight.
Security, energy and labour savings among key features
Honeywell listed several areas where the platform can support building operations.
The platform includes advanced encryption protocols designed to help protect against cybersecurity threats.
Remote diagnostics tools allow issues to be identified and addressed without requiring in-person checks.
Predictive maintenance capabilities notify operators of potential equipment issues before a failure occurs.
Energy consumption and emissions can be tracked through dedicated energy management modules.
Honeywell said these features are designed to help reduce overall costs and support operational continuity.
US buildings among first to adopt the technology
Verizon Communications has begun using Honeywell Connected Solutions in parts of its global infrastructure.
Honeywell stated that Verizon is using the system to detect critical issues before they escalate into costly failures.
Vanderbilt University has also integrated the platform in nearly 10% of its campus buildings.
The university aims to improve the performance of its older buildings while reducing energy use and improving the user experience.
Alex Kohnen, chief facilities officer at Vanderbilt University, said: “Honeywell’s Connected Solutions is an innovative platform offering the tools to use real-time data and AI-driven insights to help us optimize our buildings.
“Honeywell’s Connected Solutions is an innovative platform offering the tools to use real-time data and AI-driven insights to help us optimize our buildings.
“We look forward to creating a better ecosystem for our students, faculty and staff.”
Platform designed to meet rising operational demands
Billal Hammoud, president and chief executive officer of Honeywell’s Building Automation segment, commented on the platform’s development.
Hammoud said: “From aging buildings and rising downtime costs to skilled labour shortages and growing cyber guidelines, building owners and operators face a complex landscape of global trends that are constantly making operations more complex and costly.
“To address this, Honeywell’s Connected Solutions allows building managers to link critical building software, technologies and devices together to streamline management and compliance and help protect uptime.”
The company said the platform was developed in response to increasing pressure on building operators to reduce downtime and operational overheads.
Honeywell highlighted the system’s modular and scalable features as suited to both individual sites and large portfolios.
Installation speed was identified by the manufacturer as a key advantage over legacy systems.
Honeywell AI platform launches for US building management: Summary
Honeywell has launched Honeywell Connected Solutions for building management in the US.
The announcement was made by Honeywell on 10 June 2025.
The platform integrates building software and devices into one AI-enabled interface.
It is based on Honeywell Forge and offers remote monitoring, predictive maintenance and energy tracking.
The company said installation takes hours and reduces cost and disruption.
Early adopters include Verizon Communications and Vanderbilt University.
Verizon is using the system to predict system issues and reduce downtime.
Vanderbilt has installed it in 10% of its buildings.
The platform includes encryption to support cybersecurity needs.
It is aimed at facility managers, integrators and multi-site operators.
California’s wildfire chatbot prompts fire protection concerns
Wildfire AI chatbot raises public fire protection concerns
California’s Department of Forestry and Fire Protection has launched an AI-powered chatbot designed to improve public access to wildfire information and fire protection guidance.
According to Route Fifty, the tool was developed as part of California Governor Gavin Newsom’s 2023 executive order to boost government services using generative AI.
The chatbot, created by Citibot, provides answers based on content from the Cal Fire website and ReadyForWildfire.org.
Although it can correctly identify active wildfires and give general preparedness tips, testing by CalMatters found that it fails to offer consistent answers or the most current information.
These limitations have prompted researchers and technology experts to question whether AI deployment by California agencies is being appropriately evaluated.
Evaluation methods and chatbot performance under scrutiny
Daniel Ho, law professor at Stanford University, said that evaluation processes should be central to deploying AI tools in the public sector.
Ho said: “Evaluation is not an afterthought. It should be part of the standard expectation when we pilot and roll out a system like this.
“There should also be clear benchmarks created by an independent third party to measure performance.”
CalMatters found that small changes in how questions were phrased caused the chatbot to deliver inconsistent answers, especially on evacuation kit contents.
Although the bot was able to list items when asked about an “evacuation kit,” it failed to do so when other common phrases like “go bag” or “fire preparedness kit” were used.
In some instances, the chatbot linked to external webpages instead of providing direct answers, even though the relevant phrases were included on the referenced sites.
Containment status and recruitment information outdated
When asked about the Ranch Fire in San Bernardino County, the chatbot gave an update that was six days old, stating that the fire was 50 percent contained.
At the time of the query, the fire was already 85 percent contained.
A separate question about Cal Fire job vacancies also produced inaccurate results.
The chatbot stated that there were no openings, but a check on the state’s official employment site revealed two active listings.
Mila Gascó-Hernandez, research director at the University at Albany’s Center for Technology in Government, said: “If a fire is coming and you need to know how to react to it, you do need both accuracy and consistency in the answer.
“You’re not going to think about ‘what’s the nice way to ask the chatbot?’”
Limitations in emergency response information
The chatbot cannot provide reliable information about evacuation orders, according to testing.
Depending on how the question was phrased, it either returned no answer or gave an incomplete response.
In some cases, it incorrectly identified the issuing authority.
Cal Fire’s Deputy Chief of Communications, Issac Sanchez, said the system should be able to answer evacuation-related questions more reliably.
Sanchez said: “The answer should be ‘there doesn’t appear to be any evacuations associated with this incident’ if none are in place.
“It really was started with the intent and the goal of having a better-informed public about Cal Fire.
“We’re combing through the queries to improve future responses.”
Experts urge improved design before public release
Both Ho and Gascó-Hernandez emphasised that better public sector AI requires structured design and pre-launch user testing.
Ho said that defining performance benchmarks before selecting a vendor is essential.
Gascó-Hernandez added that representative user engagement before deployment improves the relevance of the information provided.
She said: “These user engagement and user experiences are very important so the citizen ends up using the chatbot.”
Cal Fire said it will continue to refine the system, with a commitment to host the chatbot until at least 2027.
Updates have already been made to expand its capability in response to user feedback.
California fire chatbot accuracy questioned by experts: Summary
The California Department of Forestry and Fire Protection has launched an AI chatbot to provide fire protection information.
It was developed under Governor Gavin Newsom’s 2023 executive order promoting AI in government services.
The chatbot uses Cal Fire and ReadyForWildfire.org content to respond to public queries.
It was developed by Citibot and will be hosted through 2027.
CalMatters tested the chatbot and found inconsistent and outdated responses.
Inaccuracies included outdated wildfire containment figures and missing job vacancy listings.
The chatbot failed to consistently answer questions about evacuation kits and orders.
Stanford’s Daniel Ho said evaluation must occur before system rollout.
Mila Gascó-Hernandez from the University at Albany said AI chatbots require accurate and consistent outputs.
Cal Fire’s Issac Sanchez confirmed the agency is reviewing user queries to improve the tool.
He said the tool was created to better inform the public about Cal Fire.
Cal Fire continues to update the chatbot based on public interactions.
Zelim wins British Safety Council award for AI health and safety system
Zelim recognised for AI system in British Safety Council awards
Zelim has been named the winner of the British Safety Council’s 2025 Health and Safety Transformation Award for its automatic man overboard detection system.
According to Zelim, its system ZOE was selected from a shortlist of ten international organisations, including Saudi Aramco, Pertamina, and the Universities Safety and Health Association.
The company stated that ZOE is designed for use in maritime, offshore and defence sectors and has been developed using over 7 million maritime images.
ZOE was created with support from the United States Coast Guard and uses machine learning to detect and track people in the water.
The company added that the system can function in poor visibility and high-motion conditions, removing the need for wearables or human confirmation.
Man overboard system designed with US Coast Guard
Zelim stated that the ZOE system integrates with existing vessel camera and sensor setups.
Once a person is detected, the system activates a series of automated safety responses.
These include location tagging, alert notifications, video logging, and a recovery checklist that supports emergency crews.
The company reported a 96.8 percent probability of detection at a distance of 300 metres.
It also stated that ZOE can differentiate between people, small craft, and marine animals such as seabirds.
ZOE is designed to maintain day and night functionality without manual oversight.
Deployment on vessels across maritime and defence sectors
Zelim reported that the ZOE system has already been deployed on a North Sea jack-up rig and a UK-flagged cruise ship.
It is also scheduled to be installed on a naval frigate.
According to the company, ZOE is being trialled with asset operators, ports, and defence agencies worldwide.
It explained that the technology aims to improve survivability in man overboard events.
The company also noted its additional potential for use in preventing collisions and unauthorised water entries.
British Safety Council outlines purpose of award
The British Safety Council explained that its Health and Safety Transformation Award is given to organisations that have used information in new ways to reduce harm.
It added that the award was sponsored by Croner-i and was part of its wider programme of recognising safety performance across sectors.
The Council stated that ZOE was chosen for the award due to its contribution to operational safety in high-risk environments.
It also noted that the shortlisted entries represented a range of industries and international submissions.
The award has been running annually and recognises changes made in health, safety or wellbeing.
Certification process and Lloyd’s Register approval
Zelim stated that ZOE currently holds Approval in Principle from Lloyd’s Register under ISO 21195.
This standard relates to the performance of man overboard detection systems.
The company added that it expects to complete full type approval in the near future.
It reported that ZOE would become the first system of its kind to hold complete certification under the standard.
The system’s approval process has included trials and third-party assessments to confirm detection capability and operational integration.
Zelim wins British Safety Council award for AI health and safety system: Summary
Zelim has won the British Safety Council’s 2025 Health and Safety Transformation Award.
The award was presented for its automatic man overboard detection system ZOE.
ZOE uses AI and a dataset of over 7 million images.
The system was developed with the US Coast Guard.
ZOE can detect people in the water with 96.8 percent accuracy from 300 metres.
It removes the need for wearables or manual confirmation.
ZOE integrates with vessel systems and automates emergency responses.
It is currently in use on offshore rigs and a cruise ship.
It is scheduled for deployment on a naval frigate.
Trials are ongoing with global maritime and defence organisations.
The award was sponsored by Croner-i.
The British Safety Council stated that ZOE was recognised for operational impact.
ZOE has received Approval in Principle under ISO 21195.
The system is expected to become the first fully type-approved MOB system.