Avoiding 1984: Considering Data Privacy Concerns In EHS Video Analytics

Avoiding 1984: Considering Data Privacy Concerns In EHS Video Analytics

Orwell’s 1984 is a cautionary tale of totalitarianism made possible through mass surveillance technology, not unlike the possibilities created by video analytics. This recent commercial innovation harnesses artificial intelligence and computer vision to interpret visual data feeds from CCTV, wearable devices and other sources. It automatically detects and recognizes hazardous objects, activities and patterns from large volumes of video data. EHS professionals can leverage this innovation to identify potential risks, hazards and safety violations in real time. Video analytics provides an unintrusive and continuous monitoring solution to build safer work environments, from exclusion zone detection to identifying workers at risk of fatigue. Although this is extremely valuable in the context of identifying safety risks, there are serious concerns over data privacy and security hindering adoption.

The benefits of video analytics for EHS are evident, but executives are hesitant to adopt this technology due to data privacy concerns over personally identifiable data. Surveillance enabled by video analytics creates the risk of data leaks – leading to sensitive information being used outside its intended purpose. Additionally, jurisdiction-specific privacy regulations over data collection, processing and storage present the risk of fines and litigation. Take the EU’s General Data Protection Regulation (GDPR), which requires a valid lawful basis to process personal data: consent, contract, legal obligation, vital interests, public task and legitimate interests. Depending on the context and lawful basis, individuals have the right to contest the use of their personal data. Navigating these requirements exposes an organization to many compliance risks and possible fines.

Collecting personal data creates the potential for further organizational challenges when creating a collaborative safety culture. While the intention is to enhance safety practices, the constant monitoring and recording enabled by video analytics could inadvertently lead to a culture where individuals are singled out and face punitive action for lapses or incidents. This undermines efforts to foster a culture of open reporting and collective responsibility. Workers may see video analytics as a barrier to effectively completing work – or doing so without future repercussions – and devise potentially more unsafe solutions to avoid discovery, instead of engaging with the broader team in proactively mitigating safety risks.

To combat this concern, EHS video analytics providers are continuously developing solutions that comply with the developing AI regulatory landscape. The core data privacy concern is capturing and processing information that can identify individuals. Recognizing this, AI-powered video analytics provider Intenseye announced a 3D anonymization technique to maintain greater employee privacy in April 2023. This uses deep learning algorithms that initially identify individuals, track their movements in 3D space and perform 3D keypoint detection to capture postures, before the 3D anonymization technique removes the individuals from the scene and renders an animation on the in-painted image. The Intenseye platform also allows users to remove timestamps from visual data to further separate the individual from the incident. Together, these features create a solution that minimizes the personal data captured and enables users to drill down on incidents while fostering a collaborative safety culture. Firms looking to implement video analytics or AI technology must handle personal data carefully and ensure they are keeping pace with the developing regulatory landscape.

To learn more about EU regulations on AI, please see this report. To read more on AI use cases for EHS, please see our report Strategic Focus: Improving Health And Safety With AI.