The World Economic Forum (WEF) has announced the release of its new facial recognition framework in an effort to drive the safe and trustworthy use of the technology.
The ‘Framework for Responsible Limits on Facial Recognition’ was devised by key industry players, policymakers, representatives from academia as well as civil society and WEF itself. They plan to use the framework to manage the risks of unethical practices or misuse pertaining to the technology and to ensure greater safety and privacy.
It has been forecasted that by 2024, the facial recognition market will be valued at $7 billion, which is double of what it is currently worth.
Kay Firth-Butterfield, Head of AI and ML at the World Economic Forum, said, “Although the progress in facial recognition technology has been considerable over the past few years, ethical concerns have surfaced regarding its limitations.”
She added, “Our ambition is to empower citizens and representatives as they navigate the different trade-offs they will face along the way.”
The framework works towards operationalizing use cases for two different audiences: policymakers and engineering teams and members from each team would play two different roles as well.
The first contributors include industry representatives such as Amazon Web Services and Microsoft, policymakers like French Parliament members and OPECST, as well as academics and civil society organizations.
The second group are observers which include the French Data Protection Authority and the French Digital Council.
The framework is made up of four steps. The first step is to define what the responsible use of facial recognition technology would entail, with principles focusing on bias mitigation, the proportional use of technology, consent, accountability, privacy, children’s rights and right to accessibility.
Following this, the next step would be to support product teams to ensure that the developments are “responsible by design”, with a focus on four dimensions. They would need to justify their use of facial recognition technology, devise a data plan which matches end-user characteristics, alleviate any room for biases and inform end-users.
Then, they would need to assess to what extent the designed system will be held responsible. This will be done through an assessment questionnaire which would outline the rules and regulation which would need to be accounted for and respected for each use case.
The last step of the framework includes validating compliance through the use of an audit framework by a trusted third party.
Computer vision researcher at the French graduate school and research center, Eurecom Sophia Antipolis, Jean-Luc Dugelay stated that recent developments in emerging technologies have enabled significant breakthroughs in areas that are very much related to facial recognition technology.
“For that reason, I believe that it is essential to accompany these advances in science with a global policy reflection on the appropriate use of this technology, through a multi-stakeholder collaboration that involves academics, engineers, technology providers and users, policymakers, lawyers and citizens,” he added.