Cognitive AI may sound like robots becoming human, but there’s a lot more to this emerging technology than your usual dystopian story — this tech is real.
We’ll be honest: we know “cognitive AI” sounds a bit too much like “human robot.”
When IBM’s Deep Blue beat Garry Kasparov in the 1997 World Chess Championships, the world at large finally saw the potential future artificial intelligence (AI) could bring us. It was also the year widespread AI panic gripped the masses, too.
But the distrust of AI technology is misplaced; even cognitive AI, which we’ll be covering in-depth today, relies on human intervention and goal-setting the start.
What is Cognitive AI?
Cognitive AI is the result of coupling two different tech fields: cognitive computing and artificial intelligence.
What is cognitive computing?
Cognitive computing is the algorithmic process a computer goes through to understand and simulate human reasoning and behavior based on data mining and pattern recognition. It sounds a lot like AI on paper, but the major difference comes in the long-run.
What is artificial intelligence?
Artificial intelligence (AI) is a specialized field of computer science that focuses on producing machines that mimic the natural intelligence of humans and animals.
Cognitive computing vs artificial intelligence
Both cognitive computing and AI are run using self-learning algorithms, but their one major differentiator is how their output is used.
In all AI, the output is “final,” per se. AI’s biggest strength is leveraging its’ algorithm to continually learn and evolve, ultimately ending up with better-than-human decision-making, at least in terms of speed.
Cognitive computing, on the other hand, goes through a very similar algorithmic process to provide context and understanding. The output is then instead a story from the data, rather than a decision.
To keep it simple, AI is designed to make decisions for you, while cognitive computing is designed to help you make the decision for yourself.
The end result of coupling these two together is a computer that not only self-teaches like a human and processes new information like a human, but can effectively contextualize, reason, and understand actual humans, too.
How does Cognitive AI work?
Cognitive AI melds the best of cognitive computing and AI by capturing and weaving together structured data into an accurate representation of human intelligence — or as close as the algorithm can get, at least.
Like all good AI or cognitive computing, cognitive AI continually develops over time, creating better data stories that end up with better decisions made. There are elements of human intelligence that AI will never be able to recreate, like imagination, for instance, and maybe it’s best kept that way.
Key attributes to cognitive AI
Naturally, cognitive AI constantly gets confused with other subsections of AI like deep learning. But just like cognitive computing is different from AI, cognitive AI is different from deep learning.
To help clear the confusion — and to improve implementation in commercial applications — the Cognitive Computing Consortium has decided the following attributes must be met in order for an algorithm to be considered cognitive:
Adaptive: When your job is to effectively tell a story using data, your job never ends. Stories change as new information is received — and the data pipeline doesn’t stop. Cognitive AI must be adaptive to new data in order to tell the most accurate stories.
Interactive: The whole point of developing cognitive AI is to help people make better decisions, and we can’t do that if we can’t understand the data we’re seeing. Cognitive AI must be user-friendly. Note: “User-friendly” doesn’t necessarily mean constant human interventions. Cognitive AI should still “think” for itself without human input.
Iterative and stateful: If a problem is unsolved, cognitive AI won’t just fill in the gaps — it’ll draw on more data to ensure the most accurate representation of your available solutions.
Contextual: Contextual data like time, location, user details, syntax, culture, tasks, requirements, and goals all inform cognitive AI’s process. Not just the output, either, but where data is drawn from and how it’s presented to the user is all informed by context.
Where you’ll find cognitive AI today
The State of Georgia Government Transparency and Campaign Finance Commission has employed cognitive AI to help with their endless campaign finance disclosures workload. Every month, the department processes over 40,000 disclosures — many of which are handwritten. The cognitive AI is able to read and digitalize the notes faster than any human could, allowing the team to focus on the actual task at hand: Resolving the disclosure.
Many hospitals around the world are beginning to explore artificial intelligence as a way to speed up diagnosing a patient. Given that all diagnoses are a story based on the account of the patient, cognitive AI’s capabilities perfectly match the contextual needs a patient has and the vast database of knowledge medical school teaches you to result in higher first-time diagnosis rates.
While the public’s got their attention on public self-driving cars, dangerous driving industries like mining and trucking are adopting cognitive AI autonomous vehicles to drive the long or dangerous stretches humans have had to in the past.
One of cybersecurity’s biggest problems (or benefits?) is the nonstop onslaught of new hacking tactics utilized by malicious internet users. Malware, viruses, and Anonymous are constantly reinventing their strategies, making cognitive AI’s situational awareness the best vanguard for any individual or company serious about security.
Cognitive AI is shaping AI’s future
If the human experience “depends on everything that can influence states of the human brain,” (Sam Harris) then cognitive AI is our best representation of the human experience so far. We live in our own individual databases; the context always supplies truth behind our reactions to everything — even things that haven’t happened yet.
Cognitive AI makes the future of AI seem incredibly optimistic — especially in recognizing, perceiving, translating and recommending ways we can improve our daily lives. And ultimately, that’s what AI is here for: To improve our lives.
If that’s a future worth being scared of, who knows how long it would’ve taken us to land on the moon?