Strengthening HSSE through Vision-AI

Every year, our industry experiences serious incidents linked to hazards we know well. Line-of-fire exposures remain one of the leading causes of fatal and permanent impairment injuries, while recurring incidents continue to highlight challenges in the consistent application, verification, and effectiveness of critical controls associated with the IOGP Life-Saving Rules.

The challenge is rarely a lack of standards or procedures. More often, it lies in our ability to recognise when controls are weakening, barriers are becoming less effective, or risk is increasing in everyday operations before somebody gets hurt.

This is where Vision-AI is beginning to make a real difference.

IOGP has published Vision-AI Guidelines for HSSE Applications (IOGP Report 816), the first industry-developed framework designed to help organisations implement Vision-AI responsibly, consistently, and effectively.

Developed through the Digitally Enabled Safe Sites Initiative (DESSI), a joint initiative of the IOGP Safety Committee and Digital Transformation Committee, the guideline brings together practical experience from operators, contractors, safety professionals, and digital specialists who have already deployed these technologies across a range of oil and gas environments.

Turning everyday observations into opportunities to prevent harm

Vision-AI uses computer vision and machine learning to analyse visual information from cameras and identify conditions, behaviours, and deviations that may indicate increasing risk. Examples include missing PPE, red-zone intrusions, line-of-fire exposures, blocked escape routes, unsafe interactions between vehicles and pedestrians, and other visible signs that barriers may be weakening.

What makes the technology powerful is not simply its ability to detect these conditions, but its ability to do so consistently and continuously across large and complex operations. Unlike periodic inspections or observations, Vision-AI can help organisations identify patterns and exposures that might otherwise go unnoticed.

Early experience across IOGP member companies shows that these insights are already helping organisations strengthening their safeguards.

More than technology

One of the most important messages emerging from the work is that successful Vision-AI deployment is not fundamentally a technology challenge.

While the technology itself continues to evolve rapidly, experience across member companies showed that the determining factors for success are often governance, workforce trust, leadership behaviours, data management, and alignment between safety, operations, and digital teams.

For this reason, Report 816 is not a technical specification. It is an implementation framework that helps organisations answer critical questions:

• What problems are we trying to solve?
• Which use cases create the greatest safety value?
• How should success be measured?
• How do we build workforce trust?
• How do we ensure AI is used responsibly and ethically?
• What governance is needed to sustain performance over time?

Addressing these questions early helps organisations implement Vision-AI consistently, ensuring effort is directed towards improving control effectiveness and reducing the potential for serious harm.

A common language for the industry

Another major benefit of the guideline is standardisation.

As Vision-AI adoption has accelerated, companies and vendors have developed different terminologies, classifications, and reporting approaches. This makes it difficult to compare performance, share lessons learned, or understand which solutions are delivering the greatest value.

IOGP Report 816 establishes a common taxonomy and minimum data structure for Vision-AI observations across seven HSSE domains, creating a shared language that can support benchmarking, learning, and wider industry adoption.

For Members, this means less time reinventing approaches and greater confidence that solutions are being implemented using a framework developed collectively by industry practitioners.

People at the centre

Perhaps most importantly, the guideline makes clear that Vision-AI should be deployed as a tool to support people, not monitor them.

The framework explicitly positions Vision-AI as an enabler of safer work rather than a surveillance mechanism. It incorporates human performance principles, responsible AI requirements, privacy considerations, and guidance on leadership behaviours that encourage learning rather than blame.

When introduced in this way, Vision-AI can help organisations have better coaching conversations, reinforce positive behaviours, and gain a richer understanding of how work is actually performed.

Value for Members

Report 816 provides Members with a practical roadmap for adopting one of the most rapidly evolving technologies in the safety landscape.

By bringing together operational experience, governance principles, ethical safeguards, and a common industry framework, the guideline helps Members deploy Vision-AI with greater confidence, accelerate implementation, reduce duplication of effort, and focus attention where it matters most: identifying risk earlier and preventing fatalities and permanent impairments.

Vision-AI Guidelines for HSSE Applications (IOGP Report 816) is available from the IOGP Publications Library and Members’ Area.