Despite nearly 60 years of experimentation with artificial general intelligence (AGI), mankind still hasn’t achieved the goal of creating a single machine that can perform all cognitive tasks as well as humans can.
But what if there was a way to attain AGI-level artificial intelligence (AI) using today’s technology?
The fact is that artificial narrow intelligence (ANI) systems are already available that can approximate the capabilities of artificial general intelligence by employing multiple cognitive engines in a parallel-processing architecture. And with the number of commercially-available cognitive engines expanding exponentially, this approach will enable ANI to rapidly overtake and even surpass human capabilities. Moreover, with the AI market set to expand by nearly a factor of 60 from 2016 to 2025, this method represents the best way for companies to capitalize on the industry’s growth potential.
Chasing the dream of artificial general intelligence
Since the dawn of AI technology, academics and researchers have aspired to develop AGI in the form of one machine that can master multiple, supremely difficult tasks. These capabilities include emotional response, moral decision making and understanding the nuances of interpersonal communication. While many firms continue to pursue achievements in AGI, the construction of such a machine could still be as much as decade away from reality.
Narrowing down the challenge
However, the more practical application of ANI—i.e., the training of software to do one particular thing extremely well—is already prevalent. It’s estimated that there are more than 5,000 ANI algorithms available today, with that number set to rise to the millions during the next five years.
ANI can eschew the concept of a single all-encompassing artificial general intelligence model while still producing AGI-like results by applying a collection of specialized algorithms via parallel processing then combining the results. These algorithms can perform functions including transcription, object detection, sentiment analysis and facial recognition. The indexed output of these various cognitive engines then can become searchable, organizable, shareable and then merged to become the basis for rich analytical applications.
This unlocks significant value from audio and video media, turning artificial intelligence into actionable intelligence. Companies of all sizes and in all industries can have the flexibility to tailor their individual business requirements for output, budget, accuracy and processing speed. This type of software platform can be cloud-based and driven by applications programming interfaces (API), or it can be amalgamated as a local on-premise solution. It also can use a blend of both models to meet the privacy and security requirements of any organization.
AI in the app economy
The key to realizing this vision will be the development of an open AI ecosystem that makes the vast array of cognitive engines available in a single online marketplace. Such a marketplace could yield lucrative business opportunities for algorithm developers, application developers and end users.
Much in the way that Salesforce.com’s AppExhange and Apple’s App Store spurred significant investment in software development and led to the rise of the app economy, the establishment of a structured marketplace for AI will be key to facilitating widespread custom development based on a transparent and well-defined economic model.
Building an AI business for today’s world
While AGI remains a potential technology for tomorrow, ANI can deliver on the promise of AI today. This can be accomplished by aggregating cognitive computing algorithms into a single software-as-a-service offering, extending that platform to a variety of computing environments and developing a transparent app ecosystem. With this approach, it’s possible to create AI technology with human-like capabilities—and to build an AI business that can quickly expand into a multi-billion-dollar opportunity.
Tyler Schulze is vice president, strategy & development at Veritone. He serves as general manager for developer partnerships, cognitive engine ecosystem, and media ingestion for the Veritone platform. Learn more about our platform and join the Veritone developer ecosystem today.