If you develop AI systems for structural integrity monitoring, non-destructive testing analysis, risk-based inspection, or fitness-for-service assessment of pressure equipment placed on or operated in the EU market, Article 111 of the EU AI Act directly governs your compliance obligations. EU AI Act Art.111 amends Directive 2014/68/EU — the Pressure Equipment Directive (PED) — to formally integrate EU AI Act requirements into the pressure equipment conformity framework, extending the high-risk AI pathway to AI-enabled inspection systems, structural health monitors, and safety-critical integrity assessment platforms serving refineries, power plants, chemical facilities, and pipeline operators.

The amendment follows the same Annex I bridge mechanism as Art.104 (machinery), Art.105 (agricultural vehicles), Art.106 (two/three-wheel vehicles), Art.107 (motor vehicles), Art.108 (marine equipment), Art.109 (radio equipment), and Art.110 (batteries): AI systems associated with pressure equipment that qualify as safety components, integrity monitors, or protective controllers under Directive 2014/68/EU now sit within EU AI Act Annex I, activating the Art.6(1) high-risk pathway and triggering the full Title III EU AI Act compliance stack. For pressure equipment AI developers, inspection service companies, and industrial operators, this creates dual obligations spanning both PED conformity and EU AI Act technical documentation, QMS, conformity assessment, and post-market monitoring.

Art.111 introduces a compliance paradigm absent from all other amendments in the Art.104–112 series: the primary AI applications under Directive 2014/68/EU are not embedded in new products placed on the market, but applied to pressure equipment already installed and operating. In-service inspection AI, risk-based inspection scheduling platforms, and fitness-for-service assessment tools are retrofitted safety layers governing equipment that may have been installed decades before the EU AI Act entered force. This in-service inspection dimension creates a compliance timeline and actor distribution fundamentally different from the vehicle and electronics-sector amendments preceding Art.111.

What Directive 2014/68/EU Covers

Directive 2014/68/EU on the harmonisation of the laws of the Member States relating to the making available on the market of pressure equipment governs equipment designed and built to operate with maximum allowable pressure PS greater than 0.5 bar gauge. The directive covers a broad range of industrial and commercial equipment:

Pressure vessels: Closed containers designed to contain fluids — gases, liquids, or vapour — at pressures above atmospheric. Examples include reactor vessels in chemical plants, heat exchangers in refineries, storage tanks for liquefied gases, and autoclave systems in pharmaceutical manufacturing.

Piping: Assemblies of pipe components intended to convey fluids, including pipework, pipe fittings, tubing, tube fittings, expansion joints, and associated valves. Industrial process piping in petrochemical plants and steam distribution networks are primary examples.

Steam generators: Boilers for generating steam for power generation, industrial process heating, or district heating. Water-tube boilers in power stations and shell boilers in industrial facilities are covered.

Safety accessories: Devices protecting equipment against exceeding allowable limits — pressure relief valves, rupture discs, safety shut-off valves, and limiting sensors.

Pressure accessories: Devices with operational functions required for pressure equipment operation — valves, regulators, flow controllers, and instrumentation with pressure-containment functions.

The directive applies a four-category classification scheme based on the product of maximum allowable pressure (PS) and volume (V) or nominal size (DN), with higher categories indicating higher hazard potential:

AI systems embedded in or associated with Category III and IV pressure equipment — where third-party notified body involvement is mandatory — are the primary subjects of Art.111's high-risk classification pathway.

What Art.111 Actually Changes

Article 111 is a targeted legislative amendment inserting EU AI Act obligations into the Directive 2014/68/EU conformity framework. Like Art.104–110, it operates through the Annex I bridge mechanism.

EU AI Act Annex I lists Union harmonisation legislation whose product scope intersects with the high-risk AI classification pathway. Art.111 adds Directive 2014/68/EU to that list. AI systems that function as safety components, integrity monitoring systems, or protective controllers within or for pressure equipment covered by the PED become Annex I pressure equipment AI — and therefore follow Art.6(1)'s high-risk classification rule when performing a safety function within or for the pressure equipment.

Three conditions determine whether a pressure equipment AI system becomes high-risk under Art.111:

1. The associated equipment is pressure equipment within the scope of Directive 2014/68/EU (PS greater than 0.5 bar)
2. The AI system functions as a safety component of that pressure equipment, or performs structural integrity monitoring, inspection analysis, or fitness-for-service assessment for Category III or IV pressure equipment
3. The pressure equipment is required to undergo third-party conformity assessment under the PED or applicable sector legislation

Where all three conditions are met, the developer faces dual compliance: the PED conformity process plus the full EU AI Act Title III stack.

The In-Service Inspection Paradigm: Art.111's Unique Angle

Art.111 introduces a compliance dimension absent from all preceding amendments in the Art.104–112 series. For vehicles (Art.105–107), marine equipment (Art.108), radio equipment (Art.109), and batteries (Art.110), the primary AI compliance event is at market placement — when the product containing AI is first put on the EU market and undergoes initial conformity assessment.

For pressure equipment AI under Art.111, the dominant AI applications are in-service inspection systems applied to equipment already installed, certified, and operating. Pressure vessels, boilers, and industrial piping are designed for service lives measured in decades. A reactor vessel installed in 1998 may have an expected service life to 2030 and will be inspected repeatedly during that period using AI-enhanced inspection tools that did not exist when the vessel was designed.

This creates three Art.111-specific compliance challenges:

Retrofitted AI safety layers: An AI structural integrity monitoring system added to an existing certified pressure vessel does not go through the PED conformity assessment of the vessel itself. It is a new AI safety component applied to existing infrastructure. Its EU AI Act compliance status depends on whether it performs a safety function — protecting against failure — for Category III/IV pressure equipment.

Inspection AI used by notified bodies: AI systems used by notified bodies and inspection authorities to conduct PED conformity assessments — AI-assisted radiographic film reading, AI ultrasonic defect characterisation, AI weld quality classification — create a novel compliance loop. The AI tool used to conduct the conformity assessment is itself an AI system that may require EU AI Act compliance if it performs a material safety determination.

Fitness-for-service determinations: AI systems implementing API 579-1/ASME FFS-1 fitness-for-service methodology — assessing whether a crack, corrosion thinning area, or deformation in operating pressure equipment is below critical flaw size and safe to continue operating — make a direct safety determination for live industrial infrastructure. These systems are definitively high-risk regardless of whether the pressure equipment itself is newly placed on the market.

Definitively High-Risk Pressure Equipment AI Systems

AI Structural Integrity Monitoring Systems for Refineries and Chemical Plants

Continuous AI-driven structural integrity monitoring systems for operating pressure vessels in refineries and chemical plants are definitively high-risk. These systems integrate sensor data — strain gauges, temperature sensors, corrosion monitoring probes, accelerometers — with AI models trained on pressure vessel failure modes to provide real-time integrity status and failure probability estimates.

Refineries operate Category III and IV pressure equipment — hydrocracker reactors, crude distillation towers, reformer heaters, heat exchangers — under high temperatures and hydrogen service conditions. AI monitoring systems that detect hydrogen-induced cracking, creep damage, or high-temperature corrosion in these vessels provide the primary warning of impending structural failure. A missed alert or false all-clear from an AI integrity monitor for a high-pressure hydrogen reactor could precede a catastrophic rupture with potential for fatality, explosion, and release of toxic or flammable process fluids.

Providers operating in this space include Eddyfi Technologies (AI-enhanced electromagnetic testing platforms), MISTRAS Group (acoustic emission monitoring with AI signal classification), Olympus (AI ultrasonic testing with automatic defect sizing), and Aucerna (AI asset integrity management). Their systems embedded as the primary integrity monitoring layer for Category III/IV vessels are Art.111 high-risk AI systems requiring the full Title III EU AI Act compliance stack.

AI Non-Destructive Testing Analysis Platforms

AI systems that automatically analyse non-destructive testing (NDT) data — classifying defects, sizing indications, and determining reportability — are definitively high-risk when used for Category III and IV pressure equipment inspection. NDT methods affected include:

AI NDT analysis platforms used by inspection service companies (Bureau Veritas, SGS, Applus+, TÜV Rheinland, Intertek) to produce official inspection reports for notified body review are definitively high-risk. The AI output directly determines whether pressure equipment is certified as safe for continued operation or recommended for repair or decommissioning.

AI Fitness-for-Service Assessment Systems

AI implementations of the API 579-1/ASME FFS-1 fitness-for-service standard are definitively high-risk. FFS assessment determines whether a known flaw — a crack found during inspection, a corrosion thinning area identified by UT, or a dent or gouge on a pipeline — is below the critical flaw size for continued safe operation at full rated pressure, or requires immediate repair, reduced pressure operation, or decommissioning.

An AI FFS system that incorrectly evaluates a crack as subcritical when it is in fact at or near fracture-critical conditions for the operating temperature and pressure permits continued operation of a vessel or pipeline that is at imminent risk of sudden failure. For large pressure vessels containing flammable hydrocarbons, toxic gases, or high-energy steam, sudden failure is catastrophic.

AI FFS systems provided by engineering software companies (Codeware INSPECT, COMPRESS, Netica), inspection service firms, and industrial AI startups entering the structural integrity space are definitively high-risk AI systems under Art.111.

Conditionally High-Risk Pressure Equipment AI Systems

AI Risk-Based Inspection Scheduling Platforms

AI platforms implementing risk-based inspection (RBI) methodology — API 580/581, EN 16991 — to schedule inspection intervals for pressure equipment are conditionally high-risk. RBI uses AI to estimate the probability of failure (PoF) and consequence of failure (CoF) for each pressure equipment item, and uses the product of these to prioritise inspection resources toward highest-risk equipment.

The high-risk condition is whether the AI RBI output directly determines inspection intervals for Category III/IV equipment without mandatory engineering review. In facilities where an AI RBI platform sets inspection intervals autonomously — extending inspection intervals beyond regulatory minima for equipment AI classifies as low-risk — and the AI assessment replaces rather than supplements human engineering judgment, the system is high-risk. Where AI RBI provides input to an engineer who makes the final interval determination, it is conditionally high-risk depending on the degree of human oversight actually exercised.

AI Corrosion Monitoring and Prediction Systems

AI systems monitoring and predicting corrosion rates in chemical processing pipelines, heat exchangers, and storage tanks are conditionally high-risk. The condition is whether the AI corrosion prediction directly determines maintenance or operational decisions for safety-critical pressure equipment without independent verification. Online corrosion monitoring using electrical resistance (ER) probes, linear polarisation resistance (LPR) sensors, or ultrasonic thickness measurement, combined with AI prediction of remaining useful life before minimum allowable thickness is reached, provides real-time wall-loss tracking.

AI Pressure Relief Valve Inspection Systems

AI systems analysing pressure relief valve (PRV) performance data — set pressure drift, opening and closing characteristics, seat leakage — to predict valve failure and optimise testing intervals are conditionally high-risk. PRVs are the last line of defence against overpressure events in pressure systems. AI PRV inspection optimisation systems that extend testing intervals beyond regulatory requirements based on AI condition assessment are high-risk when applied to valves protecting Category III/IV pressure systems.

Provider-Deployer Split in Pressure Equipment AI

Compliance responsibility in the pressure equipment AI ecosystem is distributed across inspection technology vendors, engineering service companies, operating companies, and notified bodies.

NDT and Inspection Technology Vendors as Providers: Companies supplying AI-enhanced inspection hardware and software — Eddyfi Technologies, Olympus, MISTRAS Group, Baker Hughes (Waygate Technologies), and GE Inspection Technologies — are EU AI Act providers when their AI defect detection and sizing software performs safety determinations for Category III/IV equipment. They must maintain technical documentation, implement QMS under Art.17, and ensure their AI inspection platforms undergo conformity assessment before use on regulated pressure equipment.

Inspection Service Companies as Both Providers and Deployers: Third-party inspection companies — Bureau Veritas, SGS, TÜV Rheinland, Applus+, Intertek — occupy a dual role. When they develop proprietary AI inspection analysis tools (automated weld defect classification, AI corrosion mapping), they are providers. When they deploy AI tools developed by NDT equipment vendors as part of their inspection services, they are deployers with Art.26 obligations: verifying the AI tool has appropriate conformity documentation, maintaining use logs, and reporting anomalous AI performance to the provider.

Industrial Operators as Deployers: Shell, BP, TotalEnergies, BASF, Dow, Bayer, and other operators of pressure equipment in refineries, chemical plants, and power stations are deployers of pressure equipment AI. Their Art.26 obligations include: ensuring AI inspection systems used on their equipment carry appropriate conformity documentation, maintaining logs of AI inspection outputs, implementing human oversight mechanisms before AI FFS determinations affect operational decisions, and reporting AI-related near-misses or anomalies to system providers.

Notified Bodies in the AI Ecosystem: Notified bodies (NBs) that use AI tools to assist in PED conformity assessment — for example, AI-assisted review of weld radiographs or AI-automated review of hydrostatic test data — occupy a novel compliance position. The NB deploying AI in its conformity assessment work is subject to EU AI Act deployer obligations for the AI tools it uses. The quality and reliability of AI-assisted NB assessment becomes part of the PED conformity framework.

CLOUD Act Intersection for Pressure Equipment AI

Pressure equipment inspection data — NDT scan files, structural integrity calculations, corrosion measurement histories, inspection reports — is increasingly stored and processed in cloud infrastructure, creating CLOUD Act exposure for EU industrial operators.

Inspection Data in US Cloud: Major inspection service companies and industrial operators use AWS (Amazon), Microsoft Azure, and Google Cloud Platform as their primary data infrastructure. MISTRAS Group, Bureau Veritas, and Applus+ each have cloud-based inspection data management platforms. Shell's digital twin programme and BP's AI asset integrity initiatives are built on Azure. TotalEnergies uses a hybrid cloud model with substantial AWS exposure. Under the US CLOUD Act, US law enforcement can compel US cloud providers to produce stored data regardless of where servers are located.

Proprietary Inspection AI Models as Trade Secrets: AI models trained on decades of inspection data from specific plant types — hydrocracker reactors, ammonia synthesis vessels, pressurised water reactor steam generators — are significant commercial assets. AI providers have invested substantial effort in training defect detection models on proprietary datasets of NDT scans with confirmed defect outcomes. CLOUD Act subpoenas targeting AI providers' cloud infrastructure could expose competitor-sensitive model training datasets and inspection outcome databases.

Critical Infrastructure Data Sensitivity: Pressure vessel inspection data for refineries, chemical plants, and nuclear facilities contains information relevant to national critical infrastructure security. The location of known defects, equipment condition trends, and remaining useful life estimates for critical pressure systems are operationally sensitive. Storing this data in US-jurisdiction cloud environments creates a data sovereignty risk beyond commercial competitiveness — it affects critical infrastructure security.

Sovereign deployment pathway: Industrial operators in regulated sectors — refineries processing strategic fuel reserves, chemical plants in EU critical supply chains, nuclear facility operators — should route pressure equipment AI inference and inspection data storage through EU-jurisdiction cloud infrastructure. For operators subject to sectoral security requirements (NIS2, critical infrastructure legislation), EU-sovereign deployment may be a compliance requirement, not merely a commercial preference.

Technical Standards Intersection

Pressure equipment AI systems must be evaluated against both EU AI Act technical documentation requirements and applicable pressure equipment integrity and inspection standards:

AI systems providing safety determinations for pressure equipment must maintain technical documentation under EU AI Act Art.11 and Annex IV that cross-references applicable EN/ISO/API standards and maps the AI system's assessment methodology to the standard's flaw acceptance criteria or risk calculation framework.

Python Pressure Equipment AI Compliance Tracker

from dataclasses import dataclass, field
from typing import Literal

PEDCategory = Literal["I", "II", "III", "IV"]
AIFunction = Literal[
    "structural_integrity_monitoring", "ndt_analysis",
    "fitness_for_service", "rbi_scheduling",
    "corrosion_monitoring", "prv_inspection", "boiler_management"
]
CloudProvider = Literal["aws", "azure", "gcp", "eu_sovereign", "on_premise"]

@dataclass
class PressureEquipmentAISystem:
    name: str
    ped_category: PEDCategory
    ai_function: AIFunction
    cloud_provider: CloudProvider
    replaces_human_judgment: bool
    notified_body_required: bool
    in_service_application: bool = True
    used_by_inspection_body: bool = False

class PressureEquipmentAIComplianceTracker:
    def __init__(self):
        self.systems: list[PressureEquipmentAISystem] = []

    def add(self, s: PressureEquipmentAISystem) -> None:
        self.systems.append(s)

    def classify(self, s: PressureEquipmentAISystem) -> str:
        # Definitively high-risk: safety-critical functions for Cat III/IV
        if s.ped_category in ("III", "IV") and s.ai_function in (
            "structural_integrity_monitoring", "ndt_analysis",
            "fitness_for_service", "boiler_management"
        ):
            return "DEFINITIVELY_HIGH_RISK"
        # Also high-risk: RBI with no human oversight for Cat III/IV
        if s.ped_category in ("III", "IV") and s.ai_function == "rbi_scheduling" \
                and s.replaces_human_judgment:
            return "DEFINITIVELY_HIGH_RISK"
        # Conditionally high-risk: lower category or indirect safety function
        if s.ped_category in ("III", "IV") and s.ai_function in (
            "rbi_scheduling", "corrosion_monitoring", "prv_inspection"
        ):
            return "CONDITIONALLY_HIGH_RISK"
        if s.ped_category in ("I", "II"):
            return "NOT_HIGH_RISK"
        return "CONDITIONALLY_HIGH_RISK"

    def cloud_act_exposure(self, s: PressureEquipmentAISystem) -> bool:
        return s.cloud_provider in ("aws", "azure", "gcp")

    def compliance_summary(self) -> dict:
        definitely = [s for s in self.systems if self.classify(s) == "DEFINITIVELY_HIGH_RISK"]
        conditional = [s for s in self.systems if self.classify(s) == "CONDITIONALLY_HIGH_RISK"]
        not_high = [s for s in self.systems if self.classify(s) == "NOT_HIGH_RISK"]
        cloud_risk = [s for s in self.systems if self.cloud_act_exposure(s)]
        nb_overlap = [s for s in self.systems if s.used_by_inspection_body]
        return {
            "total": len(self.systems),
            "definitely_high_risk": len(definitely),
            "conditionally_high_risk": len(conditional),
            "not_high_risk": len(not_high),
            "cloud_act_exposure": len(cloud_risk),
            "inspection_body_ai_tools": len(nb_overlap),
            "dual_compliance_required": len(definitely) + len(conditional),
        }

tracker = PressureEquipmentAIComplianceTracker()
tracker.add(PressureEquipmentAISystem(
    name="Refinery reactor structural integrity monitor",
    ped_category="IV", ai_function="structural_integrity_monitoring",
    cloud_provider="azure", replaces_human_judgment=True,
    notified_body_required=True, used_by_inspection_body=False))
tracker.add(PressureEquipmentAISystem(
    name="PAUT weld defect AI classifier",
    ped_category="III", ai_function="ndt_analysis",
    cloud_provider="aws", replaces_human_judgment=True,
    notified_body_required=True, used_by_inspection_body=True))
tracker.add(PressureEquipmentAISystem(
    name="AI FFS crack assessment platform",
    ped_category="IV", ai_function="fitness_for_service",
    cloud_provider="eu_sovereign", replaces_human_judgment=False,
    notified_body_required=True, used_by_inspection_body=False))
tracker.add(PressureEquipmentAISystem(
    name="AI RBI scheduling platform (auto-approve)",
    ped_category="III", ai_function="rbi_scheduling",
    cloud_provider="aws", replaces_human_judgment=True,
    notified_body_required=True, used_by_inspection_body=False))
tracker.add(PressureEquipmentAISystem(
    name="Corrosion prediction for process piping Cat II",
    ped_category="II", ai_function="corrosion_monitoring",
    cloud_provider="azure", replaces_human_judgment=False,
    notified_body_required=False, used_by_inspection_body=False))

summary = tracker.compliance_summary()
# {"total": 5, "definitely_high_risk": 3, "conditionally_high_risk": 1,
#  "not_high_risk": 1, "cloud_act_exposure": 3, "inspection_body_ai_tools": 1,
#  "dual_compliance_required": 4}

The Art.104–112 Amendment Series

Article 111 is the eighth in the Art.104–112 series of EU AI Act sector amendments, each inserting the high-risk pathway into existing EU product regulation frameworks:

The Art.111 pressure equipment amendment introduces the in-service inspection paradigm unique in the Art.104–112 series. Unlike all preceding amendments — which regulate AI embedded in products at market placement — Art.111's primary compliance landscape covers AI systems applied to pressure equipment already in service, sometimes decades after initial certification. This temporal decoupling between product conformity and AI system compliance is the defining structural characteristic of Art.111 that developers, operators, and inspection bodies must plan for ahead of the August 2026 EU AI Act application date.

25-Item Pressure Equipment AI Compliance Checklist (Art.111)

Product and System Scoping

1. Confirm the associated pressure equipment falls within Directive 2014/68/EU scope (PS greater than 0.5 bar gauge)
2. Determine the PED category (I–IV) for each pressure equipment item the AI system monitors or assesses
3. Identify whether the equipment requires notified body conformity assessment (Category III/IV) — this determines the Art.111 high-risk threshold
4. Map all AI functions associated with the pressure equipment: structural integrity monitoring, NDT analysis, FFS assessment, RBI scheduling, corrosion monitoring
5. Identify whether AI systems are applied at market placement (new equipment) or in-service (existing installed equipment) — both fall within Art.111's scope

High-Risk Classification 6. Apply Art.6(1) test: does the AI perform a safety function for Category III/IV pressure equipment? 7. Assess whether AI NDT analysis systems used by notified bodies or inspection bodies trigger high-risk classification via Art.111 8. Evaluate AI FFS assessment systems: does AI output directly determine whether pressure equipment continues operation without mandatory engineering review? 9. For AI RBI platforms: determine whether AI autonomously sets inspection intervals for Category III/IV equipment or requires human engineering approval 10. Document classification decision for each AI system with regulatory basis under both PED and the EU AI Act

Dual Conformity Assessment 11. Plan PED conformity path for AI systems embedded in new pressure equipment and identify applicable Module (B+D, B+F, G, H1) 12. For in-service AI systems: identify the applicable conformity assessment path under EU AI Act Title III 13. Establish EU AI Act technical documentation (Art.11, Annex IV) for each high-risk pressure equipment AI system 14. Implement AI risk management system (Art.9) with pressure equipment failure mode analysis: brittle fracture, fatigue, creep, SCC, HTHA 15. Define human oversight mechanisms (Art.14) for AI FFS determinations and inspection interval extensions for Category III/IV equipment

In-Service Inspection AI Governance 16. Identify all AI inspection tools used by inspection bodies performing PED conformity assessment on your equipment and verify their EU AI Act compliance status 17. Establish inspection AI output logging under Art.72 post-market monitoring: record AI defect classifications, FFS determinations, and RBI risk scores with timestamps and software version 18. Define AI model update governance for inspection AI: retrain triggers, performance drift thresholds, model version control for inspection records 19. Implement AI performance monitoring under Art.72: track rates of missed defects, false calls, and FFS assessment accuracy against ground truth (follow-up inspection, failure events) 20. Establish AI anomaly reporting chain: AI inspection anomalies or near-misses must flow from operator to inspection service company to AI system provider under Art.26(5)

Data and Cloud 21. Inventory inspection data flows to cloud platforms: identify CLOUD Act exposure for NDT scan data, structural integrity calculations, and FFS reports 22. Assess critical infrastructure data sensitivity: pressure vessel inspection data for refineries and chemical plants may require EU-sovereign storage under NIS2 or national critical infrastructure rules 23. Evaluate inspection AI training data sovereignty: AI models trained on operational inspection data from EU critical infrastructure should not reside in CLOUD Act-exposed environments 24. Map cross-border data transfer obligations for inspection data shared between EU operator, inspection service company, and AI provider 25. Register high-risk pressure equipment AI systems in EU AI Act database (Art.71) before August 2026 application deadline

See Also

• EU AI Act Art.112: Amendment to Regulation (EU) 305/2011 — AI Structural Load Assessment, Building Material Testing and BIM High-Risk Classification (Final Series Amendment)
• EU AI Act Art.110: Amendment to Directive 2006/66/EC — AI Battery Management, State-of-Health Estimation and EV BMS High-Risk Classification
• EU AI Act Art.109: Amendment to Directive 2014/53/EU — Software-Defined Radio, Cognitive Radio and AI Spectrum Management High-Risk Classification
• EU AI Act Art.108: Amendment to Directive 2014/90/EU — Marine Equipment ECDIS and Autopilot AI
• EU AI Act Art.107: Amendment to Regulation (EU) 2019/2144 — ALKS, ADAS and Motor Vehicle AI High-Risk Classification
• EU AI Act Art.106: Amendment to Regulation (EU) 168/2013 — L-Category Vehicles Motorcycle ABS and Rider AI
• EU AI Act Art.105: Amendment to Regulation (EU) 167/2013 — Agricultural Vehicles Autonomous Guidance AI