How intelligent is Artificial Intelligence when it comes to privacy?

By:Joyce Cruickshank

08 Sept, 2021

The world of data moves fast – 90% of all the world’s data was generate in the past two years alone.

The bulk of it comes from social media, digital photos, videos, customer data, and more. Consider an intelligence software that scans through customer records to pipeline insights about the customer to a salesperson or a service representative. Could we really call it an intelligence software if it also siphoned the customer’s personally identifiable information (PII) as it delivered insights to help smoothen a business process? Absolutely not. In fact, such a software could lose the business $3m+ if the agent in that scenario leaked the information, or simply wasn’t careful enough. So, what do you do? Do you stop trusting your employees? No – you simply replace the losing solution with something that’s truly intelligent.

The fact that AI is powering some of the most delicate and sophisticated workflows across industries today and in the process, creating indispensable value for businesses makes it impossible to think of a world without AI. Black-box solutions have gained a notorious reputation even at tech giants like Amazon, Google and Facebook. However, privacy concerns and data leaks are ultimately characterized by the dynamics of big data – and it is time for AI to deliver intelligent and responsible controls to assist in protecting information and frankly employees from becoming bad actors.

In fact, AI experts highlight the need for privacy as an opportunity in the AI landscape. Today, a multitude of privacy preserving mechanisms rely on AI to anticipate and detect leakages and help reinforce security perimeters and access tiers across multiple, interconnected clouds. However, there are contextual solutions too, which must learn to recognize what PII looks like, and when, which roles can do without it. If an AI solution that mines customer data from the enterprise systems, can mask PII from unmonitored contexts and help the sales or service reps to take care of their own tasks with ease, only then could it be classified as an intelligent solution. Such a solution would not only eliminate the need for executives to handle information carefully, but also help the enterprise better trust its employees to do their job.

It is time for artificial intelligence technologies to grow beyond simulations, decision support, and optimization gimmicks. AI is now a commonplace technology in the enterprise landscape. Enterprise AI solutions like Sainapse come with onboard PII detection capabilities that can adopt to varying contexts are what is needed to protect businesses. This is what enterprise AI must deliver before it can aspire to scale beyond the known and the tested. Moreover, tightening regulations around the globe, as in the guidelines of GDPR, HIPAA, PIPEDA and other privacy laws mandate the need for AI solutions to be privacy-aware, even privacy-first. If enterprise AI cannot address the issue of privacy internally, it is doomed to be classified as artificial intelligence that breaks more than it serves.


About the Author

Joyce Cruickshank

Senior technical and operations executive partnering with C-suite and leadership to drive customer retention, operational improvements, profitability and growth.