> ## Content Index
> Fetch the complete content index at: https://www.humanoidthreats.com/llms.txt
> Use this file to discover other available public pages before exploring further.

# Humanoid Robot Security Standards: ISO 25785-1, IEC 62443 and EU Rules Explained
- URL: https://www.humanoidthreats.com/humanoid-robot-security-standards/
- Published: 2026-10-08T05:59:52.000Z
- Updated: 2026-11-05T07:00:44.000Z
- Description: Humanoid robot security standards are still forming. See what ISO 10218, IEC 62443, draft ISO 25785-1 and EU law require, and the key dates to plan for.
- Author: Noa Lindqvist
- Tags: Guides, Regulation

**TL;DR**

- There is no single humanoid robot security standard yet. Buyers and makers are assembling one from ISO 10218 (industrial robots, now with a cybersecurity assessment), IEC 62443 (industrial cybersecurity) and the draft ISO 25785-1 (dynamically stable mobile robots, including humanoids).
- In the EU, the binding dates to plan around are the Cyber Resilience Act reporting duty (live since 11 September 2026), the Machinery Regulation (20 January 2027) and full CRA obligations (11 December 2027).
- The AI Act's high-risk rules for AI inside products now start on 2 August 2028, and AI in machinery has been largely moved under the Machinery Regulation instead.

**Humanoid robot security standards** are still being written, but the rules that will shape them are already on the calendar. A humanoid sold into a factory or warehouse today sits at the intersection of robot safety standards, industrial cybersecurity standards and [new EU product law](https://www.humanoidthreats.com/who-is-liable-when-a-hacked-robot-causes-harm/). This guide maps which robot cybersecurity standards apply, what each one actually asks for, and the dates that matter.

*Last updated: October 8, 2026\. This is a general overview, not legal advice.*

## Is there a security standard for humanoid robots?

Not a dedicated one. The closest is **ISO 25785-1**, "Safety requirements for dynamically stable industrial mobile robots (legged, wheeled, or other forms of locomotion) - Part 1: Robots", developed by ISO technical committee TC 299 (Robotics). According to the [ISO catalogue entry](https://www.iso.org/standard/91469.html?ref=humanoidthreats.com), it is a committee draft (ISO/CD 25785-1) whose comment period closed on July 8, 2026\. It is not a published standard yet.

Its scope fits humanoids well: robots whose balance is actively controlled and that could become unstable without power, with any number of legs or wheels, with or without arms. It is limited to industrial workplaces where the public is excluded or restricted. Consumer, medical and military robots are out of scope, and a separate Part 2 will cover the applications that integrate these robots.

ISO 25785-1 is a **safety** standard. The draft text is not public, so we cannot say how much cybersecurity it will contain. For security, today's practical answer is to combine the standards below.

## What does ISO 10218 require for robot cybersecurity?

ISO 10218 is the core safety standard for industrial robots (Part 1 for the robot, Part 2 for the robot application and cell). Its third edition was released at the start of 2025, and the biggest change for security is that the earlier 2011 edition did not mention cybersecurity at all. The revision adds a requirement to carry out a **cybersecurity assessment** and to add protective features where threats are identified, with informative references to the IEC 62443 series, IEC TR 63074 and ISO/TR 22100-4, according to an [Analog Devices engineering summary](https://ez.analog.com/ez-blogs/b/engineerzone-spotlight/posts/changes-to-the-industrial-robot-safety-standard-iso-10218?ref=humanoidthreats.com).

In the US, [ANSI/A3 R15.06-2025](https://www.businesswire.com/news/home/20250910605228/en/New-ANSIA3-R15.06-2025-American-National-Standard-for-Industrial-Robot-Safety-Now-Available-for-Purchase?ref=humanoidthreats.com) is the national adoption of ISO 10218 Parts 1 and 2, and it lists cybersecurity guidance "as part of safety planning and implementation" among its main changes.

Manufacturers are already being assessed against these clauses. In a [May 2026 announcement](https://www.sgs.com/en-in/news/2026/05/dobot-cr-30h-series-achieves-iso-10218-cybersecurity-compliance?ref=humanoidthreats.com), SGS said a DOBOT cobot series received a verification of conformity for the cybersecurity-related requirements of ISO 10218-1:2025, covering threat modeling, access control and authentication, secure communication, configuration protection, port and interface management and secure software updates.

For humanoids, ISO 10218 matters wherever a humanoid's arms do industrial manipulation work, and its cybersecurity checklist is a useful baseline even before ISO 25785-1 is finished.

## How does IEC 62443 apply to robots?

IEC 62443 is the main international series for industrial automation and control system (IACS) security. Three parts are most relevant to robot makers and integrators:

- **IEC 62443-4-1:** secure product development lifecycle requirements for the vendor.
- **IEC 62443-4-2:** technical security requirements for individual components. The [Spanish national CERT (INCIBE-CERT)](https://www.incibe.es/en/incibe-cert/blog/iec-62443-4-2-need-secure-components?ref=humanoidthreats.com) explains that it distinguishes four component types (software applications, embedded devices, host devices and network devices) and inherits its security levels from IEC 62443-3-3.
- **IEC 62443-3-3:** system-level requirements, which an integrator uses when placing robots into a plant network.

Security levels (SL1 to SL4) describe how capable an attacker the component is designed to resist, from casual misuse up to well-resourced, highly motivated attackers. The Analog Devices summary above suggests SL2 as the likely sweet spot for most industrial robots, with SL3 worth considering when the robot can be reached remotely.

Robot makers are starting to market this. On October 6, 2026, Techman Robot [announced](https://prnewswire.co.uk/news-releases/techman-robot-strengthens-cybersecurity-for-trusted-physical-ai-deployment-302899568.html?ref=humanoidthreats.com) IEC 62443-4-2 Security Level 2 certification for its TM AI Cobot S Series after an assessment by DEKRA (the certificate and its scope were not published with the release). We expect humanoid buyers to ask for the same evidence. If you need help with assessments or OT monitoring, our [security vendor directory](https://www.humanoidthreats.com/directory/) lists IEC 62443 and OT security specialists.

## What EU rules apply to humanoid robots, and when?

Three EU laws do most of the work. None is humanoid-specific, but a connected humanoid sold in the EU will usually fall under all of them.

**Cyber Resilience Act (CRA, Regulation (EU) 2024/2847).** Since 11 September 2026, manufacturers of products with digital elements must report actively exploited vulnerabilities and severe incidents through ENISA's Single Reporting Platform: an early warning within 24 hours, a notification within 72 hours, and a final report no later than 14 days after a fix is available for exploited vulnerabilities, per the [European Commission](https://digital-strategy.ec.europa.eu/en/policies/cra-reporting?ref=humanoidthreats.com). The CRA's main security-by-design obligations apply from 11 December 2027\. For robot makers, this is the first EU law with a hard clock on vulnerability disclosure, which matters given how often [fix status is unclear for robot CVEs](https://www.humanoidthreats.com/humanoid-robot-vulnerability-tracker/).

**Machinery Regulation (Regulation (EU) 2023/1230).** It replaces the Machinery Directive and becomes fully applicable on 20 January 2027 ([TUV SUD](https://www.tuvsud.com/en-us/resource-centre/blogs/testing-and-certification/new-eu-machinery-regulation-what-you-need-to-know?ref=humanoidthreats.com)). For the first time, machinery safety law explicitly covers malicious interference. As summarized by [VicOne's research lab](https://lab-r7.vicone.com/blog/when-ai-moves-why-the-eu-machinery-regulation-matters-for-physical-ai-security/?ref=humanoidthreats.com), essential requirement 1.1.9 ("protection against corruption") requires safety-relevant hardware, software and data to be protected against accidental or intentional corruption, and 1.2.1 requires control systems to withstand "reasonably foreseeable malicious attempts from third parties". The regulation also asks manufacturers to assess risks from self-evolving or autonomous behavior, which is directly relevant to AI-driven humanoids.

**AI Act.** The Digital Omnibus on AI, Regulation (EU) 2026/1744, was published on 24 July 2026 and pushed back the high-risk deadlines: 2 December 2027 for stand-alone high-risk systems and 2 August 2028 for AI that is a safety component of regulated products ([White & Case](https://www.whitecase.com/insight-alert/eu-ai-omnibus-enters-force-amending-ai-act?ref=humanoidthreats.com)). According to [Freshfields](https://www.freshfields.com/en/our-thinking/blogs/technology-quotient/eu-ai-act-unpacked-34-the-final-digital-omnibus-on-ai-key-amendments-to-the-a-102nber?ref=humanoidthreats.com), AI embedded in machinery now falls outside the direct scope of the AI Act's high-risk rules, and the Commission can instead add AI-specific requirements through delegated acts under the Machinery Regulation. For a factory humanoid, that makes the Machinery Regulation the main EU text to watch for AI safety rules.

## Humanoid robot security standards at a glance

| Standard or law           | What it covers                                                             | Security content                                         | Status (Oct 2026)                                |
| ------------------------- | -------------------------------------------------------------------------- | -------------------------------------------------------- | ------------------------------------------------ |
| ISO/CD 25785-1            | Safety of dynamically stable industrial mobile robots, including humanoids | Not public (draft)                                       | Committee draft, comments closed July 8, 2026    |
| ISO 10218-1/-2:2025       | Industrial robots and robot cells                                          | Cybersecurity assessment, refs to IEC 62443              | Published 2025; US adoption ANSI/A3 R15.06-2025  |
| IEC 62443-4-1 / 4-2 / 3-3 | Industrial cybersecurity for vendors, components, systems                  | Core of the series, SL1-SL4                              | Published; used for certification                |
| EU Cyber Resilience Act   | Products with digital elements                                             | Vulnerability reporting, security by design              | Reporting live 11 Sep 2026; full 11 Dec 2027     |
| EU Machinery Regulation   | Machinery, including robots                                                | Protection against corruption, control-system resilience | Applies 20 Jan 2027                              |
| EU AI Act (as amended)    | AI systems                                                                 | High-risk rules; machinery largely moved out             | Product-embedded high-risk rules from 2 Aug 2028 |

Outside the EU, the most important recent rule is market access rather than a technical standard: the US FCC added foreign-produced advanced robotic devices to its Covered List in July 2026\. See our [FCC robot ban explainer](https://www.humanoidthreats.com/fcc-robot-ban-covered-list-humanoid-robots/).

## What should buyers ask a humanoid vendor today?

Until ISO 25785-1 is published, ask for evidence against the standards that already exist:

1. **ISO 10218-1:2025 cybersecurity assessment** for the robot's arms and controller, or a third-party verification of conformity.
2. **IEC 62443-4-1 and 4-2** status: is the development process assessed, and at what security level are the components certified?
3. **CRA readiness:** how will the vendor notify you of actively exploited vulnerabilities, and how fast?
4. **Machinery Regulation plan:** how are safety functions protected against corruption, and how is tampering logged?
5. **Vulnerability history:** check every model against our [Humanoid Robot Vulnerability Tracker](https://www.humanoidthreats.com/humanoid-robot-vulnerability-tracker/) and ask which firmware fixes each entry.

The recent [Unitree G1 vulnerabilities](https://www.humanoidthreats.com/unitree-g1-vulnerabilities-cve-2026-76639-76640/) show why: Bluetooth setup, cloud services and AI assistant interfaces gave attackers root access, exactly the "port and interface management" and "secure update" topics these standards cover. For AI-specific risks that standards barely address yet, read our analysis of [prompt injection against robot VLA models](https://www.humanoidthreats.com/prompt-injection-robot-vla-models/). For a ready-made procurement and deployment list, use the free [Humanoid Robot Security Checklist](https://www.humanoidthreats.com/humanoid-robot-security-checklist/).

## FAQ

**Is there an ISO standard for humanoid robots?** Not a published one. ISO/CD 25785-1 covers dynamically stable industrial mobile robots, including humanoids, but it is still a committee draft. Its comment period closed on July 8, 2026.

**Does ISO 10218 cover cybersecurity?** Yes. The 2025 edition adds a cybersecurity assessment requirement and references the IEC 62443 series. The 2011 edition did not mention cybersecurity.

**Which IEC 62443 security level should a robot meet?** It depends on the risk assessment. SL2 is a common baseline for industrial components, and SL3 is worth considering for robots reachable remotely or in high-consequence settings.

**When do EU cybersecurity rules apply to humanoid robots?** CRA vulnerability reporting has applied since 11 September 2026, the Machinery Regulation applies from 20 January 2027, and the CRA's full obligations from 11 December 2027.

## Sources

- [ISO: ISO/CD 25785-1 catalogue entry](https://www.iso.org/standard/91469.html?ref=humanoidthreats.com)
- [Analog Devices: Changes to the industrial robot safety standard ISO 10218](https://ez.analog.com/ez-blogs/b/engineerzone-spotlight/posts/changes-to-the-industrial-robot-safety-standard-iso-10218?ref=humanoidthreats.com)
- [A3 via Business Wire: ANSI/A3 R15.06-2025 released](https://www.businesswire.com/news/home/20250910605228/en/New-ANSIA3-R15.06-2025-American-National-Standard-for-Industrial-Robot-Safety-Now-Available-for-Purchase?ref=humanoidthreats.com)
- [SGS: DOBOT CR 30H series ISO 10218 cybersecurity compliance](https://www.sgs.com/en-in/news/2026/05/dobot-cr-30h-series-achieves-iso-10218-cybersecurity-compliance?ref=humanoidthreats.com)
- [INCIBE-CERT: IEC 62443-4-2, the need to secure components](https://www.incibe.es/en/incibe-cert/blog/iec-62443-4-2-need-secure-components?ref=humanoidthreats.com)
- [Techman Robot: IEC 62443-4-2 SL2 certification announcement](https://prnewswire.co.uk/news-releases/techman-robot-strengthens-cybersecurity-for-trusted-physical-ai-deployment-302899568.html?ref=humanoidthreats.com)
- [European Commission: CRA reporting obligations](https://digital-strategy.ec.europa.eu/en/policies/cra-reporting?ref=humanoidthreats.com)
- [TUV SUD: New EU Machinery Regulation](https://www.tuvsud.com/en-us/resource-centre/blogs/testing-and-certification/new-eu-machinery-regulation-what-you-need-to-know?ref=humanoidthreats.com)
- [VicOne: Why the EU Machinery Regulation matters for physical AI security](https://lab-r7.vicone.com/blog/when-ai-moves-why-the-eu-machinery-regulation-matters-for-physical-ai-security/?ref=humanoidthreats.com)
- [White & Case: EU AI Omnibus enters into force](https://www.whitecase.com/insight-alert/eu-ai-omnibus-enters-force-amending-ai-act?ref=humanoidthreats.com)
- [Freshfields: The final Digital Omnibus on AI](https://www.freshfields.com/en/our-thinking/blogs/technology-quotient/eu-ai-act-unpacked-34-the-final-digital-omnibus-on-ai-key-amendments-to-the-a-102nber?ref=humanoidthreats.com)

Get the weekly Humanoid Threats Brief → [Subscribe](#/portal/signup)