Deloitte predicts more attempts to ban facial recognition in 2022
Advisory services company Deloitte Global’s TMT (Technology, Media & Telecommunications) Predictions has projected that next year could witness a series of stricter regulations on the deployment of artificial intelligence (AI) systems – which could have far-reaching consequences on facial recognition and across industries. Most discussions, however, are unlikely to reach enforced enactment in the next year.
According to an article on Deloitte’s website, some biometric systems based on AI such as facial recognition in public spaces for law enforcement and criminal justice will face increased scrutiny, especially as the European Union is considering banning the use of the technology with some exceptions. Changes could also have major implications for deployment in other sectors such as private-sector health care, education and the financial services industry.
Deloitte posits that although AI technologies will be more effective than it was five years ago thanks to faster specialized processors, more powerful software, and larger data sets; it is set to continue to resuscitate concerns because of its implications for fairness, bias, discrimination, diversity, and privacy.
While the report predicts that AI regulation is likely to intensify, there are questions however about how exactly such regulation will happen given the fact that many AI decisions are not ‘explainable.’
A lack of full understanding of why AI algorithms make the decisions they make, the report states, renders the task of effectively regulating AI technology exponentially harder than regulating the more explainable and auditable technology that has often informed decision-making in the last century.
Explaining further, the report adds that the importance of regulation for AI cannot be over-emphasized because as its use is expanding, such regulation has to take place in order to shape the negative and positive impacts AI can have on users.
While the Deloitte Global prediction holds that regulators of AI will likely crack down on instances where algorithmic bias or other issues harm different groups of people, such regulations will impact the use of facial recognition and other AI tools by different industries and functions within them to different degrees.
Deloitte sets out four possible scenarios for the next two years. Facial recognition vendors and other stakeholders could exit jurisdictions with stricter regulations, conflicting regulations could be passed in different parts of the world, a certain set of regulations may emerge as a gold standard, like Europe’s GDPR, Deloitte suggests, or AI players could come together to perform self-regulation.
In summary, the report predicts that the coming two years could roll out an entirely different scenario as far the deployment and regulation of AI is concerned.
Homomorphic encryption, which enables data to be processed without making it vulnerable by decrypting it, and federated learning are also expected to increase in adoption next year.
Meanwhile, TMT Predictions has also made a forecast on other issues including technology in general, media, telecommunications as well as women in tech.
AI | biometrics | Deloitte | explainability | facial recognition | homomorphic encryption | regulation