Surprised? I’m not. With venture capital and global AI financing skyrocketing, AI has become a huge buzzword. It’s only natural for companies to try to pass off basic data analytics as artificial intelligence. But since investing in AI can be immensely profitable—not just monetarily, but also for real-world problem solving—pinpointing genuine AI is an important task for investors and consumers alike.
Fortunately, a quick examination of what a company’s technology actually does reveals the truth. If we had to sum up what makes AI real AI, it’s this:
Legitimate AI makes iterative decisions based on proprietary data.
Proprietary data, or metadata, is internally generated within a firm. Artificial intelligence takes in data from quality outside sources and makes a decision in response to it, whether that’s a self-driving car making a turn on the road, or Google’s AlphaGo program making a move in a game. By creating its own stream of proprietary data based on its initial decision, the program can improve its decision-making process. This where the “iterative systems” come in—systems that take a set of instructions and execute them multiple times in a feedback loop.
Say, for example, the turn of the self-driving car goes smoothly, or the move in Go creates an advantage for the opponent. AI can then learn that the decision it made actually worked or didn’t work, gathering metadata from its own reaction to the input. This is how AI can get even smarter—through supervised learning, machine learning and deep learning—which leads to major competitive and technological advantages.
This is how AI get smarter—through iterative machine learning and deep learning—and this is where the real competitive, technological, and groundbreaking advantages of AI kick in. Take, for example, a program that posts on social media according to an algorithm that determined the best time to post. That program uses neither proprietary data nor machine learning. So that program can’t be considered AI.
Let’s make the difference between data analytics and AI a little clearer. What if Tesla or Google collected data to make a car that could, in a closed course, drive straight down one street, turn right at the stop sign, and then park in an empty lot? That would be automation, and not AI. But what if it’s not a closed track—what if it’s an actual road? Does the car know how to follow the car in front of it a minimum distance, adjusting its speed and breaks to maintain it? Is it able to stop at a red light? On a live track, AI can make decisions like a human can, in real time.
What about NewsRx?
A lot of computing and technology goes into NewsRx’s BUTTER, an online proactive discovery and research tool: keyword search technology, computational journalism, data analytics… and, yes, artificial intelligence.
Butter’s AI developed over the course of the 1990s, when NewsRx initially made technology that could identify quality content versus low-quality content, in hopes of saving time for writers and editors. Starting with simple automated editing software to help writers create more content, NewsRx collected proprietary data and used supervised learning over years and decades. The end result was artificial intelligence that can conduct sophisticated research from start to finish: technology that can search and aggregate sources, identify and eliminate low-quality data, and write and format a news report based on the information.
A leading AI Venture capital firm encourages asking three questions when determining if technology is authentic AI. It turns out Butter checks all three of these boxes:
1. Is it creating its own data exhaust?
Yes. Every report that NewsRx’s computational journalism writes pulls information and data from at least two to three different vetted sources. This process generates proprietary data in the making of each original article.
2. Do they have iterative technology?
Yes. NewsRx has several iterative processes, or systems that take a set of instructions and execute them multiple times in a feedback loop, making the system smarter over time. NewsRx’s technology that uses iterative technology includes content parsing and evaluation, fact checking, identifying incorrect data, and more.
3. Is it a groundbreaking solution to a real problem?
Yes. Research is incredibly expensive, time-consuming, and challenging. Butter’s proactive alerts, which deliver concise research briefs about personalized topics and industries automatically, without the researcher having to lift a finger. This technology fundamentally simplifies the way scientists and business professionals conduct research.
So in case you were wondering if NewsRx runs Butter on actual AI, the answer is yes.
As for everything else out there, best of luck.
Read more: What is Artificial Intelligence?