Get Strategic insights into your business!
We develop and deploy solutions focused on your business needs and objectives
Eleanor Brown, Director of Business Change at Aelm
In part one, we defined artificial intelligence and its three constituent steps: perception, cognition and action. Last week we explored perception and how everyday businesses can benefit from natural language processing tools and techniques.
Now it’s time to look under the hood of what’s arguably the most complex of our three steps: cognition.
As humans, we take our brains for granted. We reason, we problem solve, we pontificate, we learn – it’s all bread and butter to the brain. But it’s incredibly difficult to replicate our cogs in machine form. Some early AI developers thought that reconstructing cognition meant machines figuring out every possible outcome in a given scenario; and then selecting the optimum action.
But what if a friend asked you what he should do with his apple. Hurling it at Kermit the Frog on St Patrick’s Day wasn’t the reply he expected. Outcomes can be limitless. So rifling through every possible iteration is a dead end: it’s sluggish to do and it’s not an accurate reflection of human thought.
If the AI end game is to have humans and machines seamlessly interacting, then finding common, pseudo-emotional understanding is key. A person decides by combining experience, personal and cultural biases and heuristics. So right now, AIs are being programmed with more ‘gut feeling’ about them. And there are some neat tools and techniques out there.
Cognitive AI is making a huge difference in the world of finance. From better algorithms and effective real-time decision-making tools, big data is even unlocking credit for the world’s financially disenfranchised. AI analysis can often predict a person’s future financial behaviour better any credit agency.
Today, AI-enabled platforms can analyse information from hundreds of data points in real time. These AIs continuously apply meaning to an individual’s actions to learn their habits and preferences. With a more holistic interpretation of a person’s needs, AIs can predict not only what one might invest in or acquire, but also when, where, and even a little around why.
Like finance, various other related and symbiotic industries are handing more and more over to AI because of its ability to scrape the web and make relevant decisions. Take financial recruitment: by assessing a person’s word choice, work history and digital personality, firms can identify individuals who’ll fit before even announcing a job opening, something invaluable in such a specialised sphere.
Then there’s AI-enabled videos which interpret a person’s facial expressions and word choices. So when your new job candidate, or loan applicant, or potential date, gets in front of a video camera, the recruiter, loan agent, or psycho would-be lover can understand what you don’t say as much as what you do.
These are examples of systems that seek not only to replicate human cognition but to improve on it by removing the inherent individual or even systemic biases that often creep into decisions. Of course, some recent cases have demonstrated that improperly trained AI systems can themselves fall into the trap of making biased decisions. But new algorithms are being developed to counteract that.
Cognition? It’s complicated. Replicating a brain is. Cognition is basically programming machines to do the equivalent of think. AIs might not be ready to replace human brains en masse, but they’re evolving fast and they present an effective, efficient option. From businesses who deal in complex decision-making to those that just want to predict outcomes, cognitive AI is a friend and an ally.
Published on KNect365 on May 4th