In April 2024, researchers published a peer-reviewed analysis of OSHA Severe Injury Reports and identified 77 robot-related workplace accidents in the US since the 2015 reporting rule took effect.
Fifty-six percent were pinch and crush injuries. Forty-four percent were impact injuries. Every one of them happened in a workspace that had passed an internal safety review and met the standard at the time.
That gap between paper compliance and real-world outcome is exactly what a modern Cobot Risk Assessment is supposed to close.
The job got harder in 2025: ISO 10218:2025 absorbed ISO/TS 15066 entirely, retired the term “collaborative robot” in favor of “collaborative application,” and started a 24-month transition window that every US plant running cobots should already be inside.
| The Practitioner’s Cheat Sheet |
| ISO 10218:2025 absorbed ISO/TS 15066 in 2025, retiring the standalone collaborative-robot specification — every active US Cobot Risk Assessment file should be re-baselined within the 24-month transition window. |
| The standard formally retired the term “collaborative robot” in favor of “collaborative application,” because safety attaches to the deployed task, not the hardware. |
| A 2024 analysis of OSHA Severe Injury Reports identified 77 robot-related accidents; 56% were pinch / crush injuries and 44% were impact injuries — both are exactly what a Cobot Risk Assessment is supposed to prevent. |
| The IFR’s 2025 World Robotics report shows global cobot installations exceeding 200,000 units in 2024, growing above 22% CAGR. The cobot market is projected to triple from $3 billion in 2025 to over $17 billion by 2033. |
| Pick one of four ISO 10218:2025 safety methods per task: safety-rated monitored stop, hand guiding, speed and separation monitoring, or power and force limiting. The wrong method is the most common Cobot Risk Assessment failure. |
| A defensible Cobot Risk Assessment integrates with the wider operational risk management program and feeds the enterprise risk register, not a standalone safety binder. |
| US plants under EU export pressure also need EU Machinery Regulation 2023/1230 readiness; ISO 10218:2025 is the de facto bridge. |
This guide rewrites the standard Cobot Risk Assessment article for US risk leaders, EHS managers, and manufacturing engineers working in 2026. We replace the old ISO/TS 15066 framing with the ISO 10218:2025 workflow, walk through the four collaborative application safety methods, and tie every recommendation to active US OSHA injury patterns and the EU Machinery Regulation 2023/1230 obligations active for any plant that exports.
Figure 1. The Cobot Risk Assessment workflow under ISO 10218:2025.
What a Modern Cobot Risk Assessment Has to Cover
A modern Cobot Risk Assessment is a documented analysis of every collaborative application a plant runs, broken down by task, workspace, and operator profile, and measured against ISO 10218:2025 and the underlying EN ISO 12100 hazard methodology. It is a working file, not a one-off audit.
The asset list, the task list, and the risk register all change as cells are reconfigured, end-effectors swap, and product mixes shift.
Two design decisions separate a defensible Cobot Risk Assessment from a binder no one trusts.
First, the assessment scopes the application, not the robot. A 5 kg cobot doing assembly is a different risk profile from the same arm doing dispense or polishing. Second, the assessment names a human owner per cell — usually the cell engineer — who keeps it current as the plant evolves.
Where the Cobot Risk Assessment Sits in the Wider Risk Stack
| Layer | What it provides | How the Cobot Risk Assessment fits |
| Hazard methodology | EN ISO 12100 risk-assessment principles | Provides identify-estimate-evaluate-control flow |
| Robot safety standard | ISO 10218-1:2025 (robot) + 10218-2:2025 (application) | Direct authority for cobot risk decisions |
| Functional safety | ISO 13849 (PLs) and IEC 62061 | Validates safety-rated stop and monitoring functions |
| Workplace safety | OSHA General Duty Clause + 29 CFR 1910 Subpart O | US enforcement floor for cobot operations |
| Enterprise risk | ISO 31000:2018 + COSO ERM | Routes residual cobot risks into the enterprise register |
Why ISO 10218:2025 Changed What Cobot Risk Assessment Means
ISO 10218:2025 is the most consequential cobot safety update since the original 2011 release. The Association for Advancing Automation (A3) notes that the new standard absorbs ISO/TS 15066:2016 entirely.
Plants still using ISO/TS 15066 as their primary reference are working from a retired specification and putting their Cobot Risk Assessment paperwork at risk during audits.
What the ISO 10218:2025 Update Means for Your Cobot Risk Assessment
- Terminology shift: “Collaborative robot” is gone; “collaborative application” is the operative term. Safety attaches to the deployed task, not the arm.
- Two-part structure: ISO 10218-1:2025 governs the robot itself; ISO 10218-2:2025 governs the integrator and the application — and absorbs the old ISO/TS 15066 power-and-force tables.
- Force and pressure tables: Updated with body-region-specific limits validated against more recent biomechanical data; the old ISO/TS 15066 thresholds are no longer the reference.
- 24-month transition: Existing applications must be re-assessed against ISO 10218:2025 within the transition window. Plants delaying re-assessment carry growing audit and litigation exposure.
- Integrator accountability: ISO 10218-2:2025 makes the integrator responsible for the application risk assessment, not just the robot OEM. Most US plants take that responsibility on themselves.
The practitioner takeaway: every active US Cobot Risk Assessment file should be re-baselined against ISO 10218:2025 within the transition window.
The Association for Advancing Automation FAQ is the most accessible US-language reference; ISO 10218-2:2025 is the authoritative document. Plants reading only the OEM datasheet are missing the integrator-side obligations.
The Four Cobot Risk Assessment Safety Methods US Plants Actually Use
Once the Cobot Risk Assessment identifies the hazard and estimates the risk, the integrator picks one of four safety methods per task.
These four methods existed under ISO/TS 15066 and carry forward into ISO 10218-2:2025. Picking the wrong method is the most common Cobot Risk Assessment failure we see in US plants.
Figure 2. Four Cobot Risk Assessment safety methods compared on ease of implementation versus continuous-collaboration flexibility.
Method Comparison Table for the Cobot Risk Assessment
| Safety method | How it works | Best fit | Watch-out |
| Safety-rated monitored stop | Cobot fully stops when worker enters the collaborative space | Predictable handoff tasks; mature workforces | Throughput penalty; not true collaboration |
| Hand guiding | Worker physically guides cobot via a safety-rated control | Path teaching; manipulation of heavy parts | Requires PL d hand-guiding device + training |
| Speed and separation monitoring | Cobot slows or stops based on real-time worker proximity | Mixed traffic cells; AGV-cobot integration | Vision / lidar accuracy under variable lighting |
| Power and force limiting | Cobot designed so any contact stays below body-region thresholds | Continuous-collaboration assembly + light pick | Force tables in ISO 10218-2:2025 are stricter than legacy ISO/TS 15066 |
Power and force limiting is the most flexible method and the most over-claimed. Many cobots marketed as “safe by design” still need a Cobot Risk Assessment to validate that the specific task, end-effector, and workpiece geometry stay below body-region force limits.
We coach US plants to demand a power-and-force limiting validation report from the integrator before signing acceptance, not just the OEM safety datasheet.
How OSHA Injury Data Should Shape Your Cobot Risk Assessment
US plants that build a Cobot Risk Assessment around the actual injury data have programs that hold up under regulator scrutiny.
Plants that rely on vendor marketing have programs that look fine until something happens. The 2024 OSHA Severe Injury Reports analysis gives the empirical base.
Figure 3. The injury patterns a Cobot Risk Assessment must engineer against, drawn from US OSHA data.
Five Hazard Categories Every Cobot Risk Assessment Should Score
| Hazard category | What it looks like in the cell | Cobot Risk Assessment control |
| Pinch and crush | Worker hand caught between cobot and fixture | Fixture geometry redesign; force-limiting validation |
| Impact | Worker struck by end-effector or arm at speed | Speed and separation monitoring; PFL force budget |
| Ergonomic strain | Repetitive reach to share workspace with cobot | Workstation redesign; rotation; ergonomic review |
| Electrical and thermal | Faulty wiring; overheating end-effector | Insulation, grounding, thermal cutoffs |
| Cybersecurity | Compromised teach pendant or PLC | Network segmentation; signed firmware; OT security controls |
The cybersecurity category is the one most US Cobot Risk Assessments still understate in 2026. A cobot that fails a CISA Industrial Control Systems advisory is no longer a safe cobot. We see plants treat IT and OT as separate worlds; the next OSHA inspector will not.
Where Cobot Risk Assessment Programs Stall — And How to Unstick Them
Even with a clear standard, US Cobot Risk Assessment programs stall in predictable patterns.
The traps below appear across sectors and across plant maturities; recognizing them up front saves the inspector visit, the workers’ compensation claim, and the post-incident audit weeks of work.
| Pitfall | Root cause | Remedy |
| One assessment for the whole cobot | Risk scoped to hardware, not application | One Cobot Risk Assessment per task / end-effector / workpiece combination |
| Reliance on “safe by design” marketing | OEM datasheet treated as proof of safety | Integrator-led power-and-force limiting validation report |
| Ignoring the new ISO 10218:2025 thresholds | Plant still references ISO/TS 15066 force tables | Re-baseline the assessment within the 24-month transition window |
| No human owner per cell | Assessment owned by an absent corporate function | Cell engineer signs the file; refresh on each program change |
| Static workspace assumptions | Risk evaluated once, not after layout changes | Trigger re-assessment on cell, fixture, or end-effector changes |
| Skipping cybersecurity | OT treated as out of safety scope | Add OT cyber controls; pull from information security risk management |
| Training gap | Workers see slides once, never again | Annual refresh; live drills on emergency stop and hand guiding |
Common Cobot Risk Assessment Questions Practitioners Ask
How is the Cobot Risk Assessment different from a traditional industrial robot risk assessment?
Traditional industrial robot risk assessments under ISO 10218 v1 assume the robot operates inside a guarded fence, with no expected human entry.
A Cobot Risk Assessment under ISO 10218:2025 assumes humans share the workspace by design, so the analysis must model task-specific contact scenarios, force and pressure limits per body region, and dynamic safeguards.
The result is a deeper, application-specific assessment rather than a one-time installation review.
What standards govern a Cobot Risk Assessment in the US in 2026?
US plants build a Cobot Risk Assessment on ISO 10218-1:2025 and ISO 10218-2:2025 (which absorbed ISO/TS 15066), with EN ISO 12100 supplying the underlying hazard methodology and ISO 13849 providing functional-safety performance levels.
OSHA’s General Duty Clause and 29 CFR 1910 Subpart O provide the US enforcement floor. Exporters also need EU Machinery Regulation 2023/1230 alignment for any cobot deployed in or destined for EU plants.
How long does a thorough Cobot Risk Assessment take?
A first-time Cobot Risk Assessment for a single new application typically takes two to four weeks of focused work, depending on task complexity, end-effector design, and workspace integration.
Plants running a fleet of cobots can amortize the effort by reusing scenario libraries and force-limiting validation reports across similar applications. Annual reassessments and triggered reviews after cell changes typically run three to seven days each.
Who should sign off on a Cobot Risk Assessment in a US manufacturing plant?
A defensible Cobot Risk Assessment usually has three signatures: the cell engineer (technical accuracy), the EHS manager (workplace safety alignment), and a senior operations leader (business acceptance of residual risk).
The integrator’s signature on the validation report sits underneath the plant’s three-signature approval. Auditors and OSHA inspectors look for this layered ownership; a single signature is a flag, not a control.
Are power-and-force-limiting cobots automatically safe without a Cobot Risk Assessment?
No. Power-and-force-limiting cobots are designed to keep contact below body-region thresholds, but the Cobot Risk Assessment must validate that the specific task, end-effector, workpiece, and workspace stay within those limits.
Sharp tools, heavy payloads, pinch geometries, and unexpected operator postures can all push contact above thresholds. Vendor marketing language is not a substitute for an integrator-led validation report.
What changed in the Cobot Risk Assessment force-and-pressure limits under ISO 10218:2025?
ISO 10218-2:2025 absorbed and updated the body-region force and pressure tables originally published in ISO/TS 15066:2016.
The new limits reflect more recent biomechanical research and apply per body region (head, torso, hands, fingers, etc.).
Plants relying on the legacy ISO/TS 15066 tables should expect different — usually stricter — thresholds in the updated standard, and should re-validate any power-and-force-limiting application within the transition window.
How does cybersecurity fit into the Cobot Risk Assessment?
A modern Cobot Risk Assessment treats cybersecurity as a safety control. Compromised teach pendants, unsigned firmware, and unsegmented OT networks can all force a cobot into unsafe behavior.
CISA Industrial Control Systems advisories now flag cobot vulnerabilities directly. US plants should add network segmentation, signed-firmware enforcement, and authentication on the teach interface to the Cobot Risk Assessment file alongside the physical safeguards.
Does a small US manufacturer really need a formal Cobot Risk Assessment?
Yes. The OSHA General Duty Clause applies regardless of plant size, and a documented Cobot Risk Assessment is the standard evidence that the employer has met the duty.
Smaller US manufacturers can run a credible assessment with a single cell engineer plus an outside reviewer and finish the first cell in two to three weeks. The cost is a fraction of one workers’ compensation claim, before counting reputational damage.
Where Cobot Risk Assessment Is Heading: 2026-2028
Cobot Risk Assessment practice is shifting from a static document to a continuous control loop, and three different pressures are driving the change at once.
The most immediate is the ISO 10218:2025 transition window itself: plants that re-baseline early will set the de facto industry benchmark, and plants that delay will find their old ISO/TS 15066 paperwork being challenged in regulator inspections and customer audits through 2027.
Then there is AI-driven cobot programming, which is the wild card for 2026. Generative tools that auto-generate motion programs from natural-language briefs are arriving fast, which moves the safety question from “is this taught path safe?” to “is this generated path safe?”
The Cobot Risk Assessment will need a new validation step for AI-generated motions, including model-card review and override documentation.
Cybersecurity convergence is the pressure most plants still underestimate. CISA advisories on industrial control systems now routinely flag cobot platforms, and the SEC cybersecurity disclosure rule is beginning to surface OT events at public US manufacturers.
Expect Cobot Risk Assessment files to merge with OT cyber risk registers by 2027, with one register feeding both safety and cyber boards.
Ready to Operationalize Your Cobot Risk Assessment Under ISO 10218:2025?
At riskpublishing.com we help US manufacturers, EHS managers, and risk leaders re-baseline their Cobot Risk Assessment files against ISO 10218:2025, calibrate force-and-pressure limits to specific applications, and integrate cobot risk into the wider operational and cyber risk register.
Practical deliverables include the assessment template, hazard scenario library, integrator validation checklist, and a litigation-ready documentation pack.
Explore our risk advisory services, or contact us to scope a Cobot Risk Assessment review tailored to your plant footprint, regulatory exposure, and 2026-2028 cost-containment targets.
Related reading on riskpublishing.com: operational risk management, a guide to risk assessment methodology, how to conduct a risk assessment, approaches and tools for risk identification, the definition of hazard and risk assessment, operational risk management framework, and scenario based risk assessment.
Chris Ekai is a Risk Management expert with over 10 years of experience in the field. He has a Master’s(MSc) degree in Risk Management from University of Portsmouth and is a CPA and Finance professional. He currently works as a Content Manager at Risk Publishing, writing about Enterprise Risk Management, Business Continuity Management and Project Management.