Looking at the current state of artificial intelligence, you could be forgiven for thinking that Clarke was absolutely right. Most machine-learning tools evolve independently as they’re trained, and frequently the end result is an AI algorithm that operates within a “black box,” spitting out results based on a process that nobody — not the designer, and certainly not the end-user — really understands. The algorithm simply works, as if by magic.
Here at Dathena, we believe that’s simply not good enough. When it comes to security, a black-box solution is the equivalent of simply saying “Hey, trust me!” As any security expert or data-privacy officer will tell you, blind trust has no place in a rigorous security system.
We’re big fans of AI and machine learning, and we’re using those technologies to develop new ways to help companies to identify, analyze, and protect private information. According to Gartner: “Security and risk management leaders need to include artificial intelligence applications for rapid, inclusive and consistent support for compliance insight, vast and continuous data discovery, subject’s rights management, etc.”
But we also firmly believe that any security system you don’t understand completely is a potential vulnerability. In creating the AI tools that underpin the Dathena Privacy platform, we’ve developed tools that work incredibly well, delivering fast, cost-effective, and data-driven insights that allow companies to protect private information far more effectively. But we’ve done so while simultaneously ensuring that all our AI solutions are both explainable and deterministic.
“Explainability” does just what it says on the tin: it means that all the results our platform delivers spring from methods and processes that can be clearly articulated. We can tell our customers exactly how any given result was generated, and we can help them understand exactly what data was used, and what calculations were performed, to generate any given output.
And our solutions are also “deterministic,” which means they are based on clear logical rules and relationships between established facts. There are no heuristics or stochastic elements introducing uncertainty or randomness into our algorithms: run the same process a second or third or fiftieth time, and you’ll get the same result every time.
The combination of those two things means that Dathena Privacy — almost uniquely in the AI space, it sometimes seems — is comprehensible by design. When you’re using our platform, you’ll never feel like you’re watching a magic trick unfold before your eyes — you’ll simply know you’re using a reliable, well-designed tool that automates makes your data-security processes exponentially easier to manage.
That might feel like a game-changer, but we promise it’ll never feel like magic. Because when it comes to AI, only technology that can’t be mistaken for magic is sufficiently advanced to be relied upon to protect your customers’ private information.