With the introduction to the public and mass-marketing push behind IBM’s Watson artificial computing platform, the world of AI and intelligent learning machines has officially arrived. Of course, Watson’s not the only game in town. Multiple tech conglomerates and startups alike are already achieving amazing results, but many efforts are focused on improving Big Data applications and enterprise-level operations, leaving smaller companies wondering how – and if – AI will show promising returns.
What Can AI Do Right Now?
Artificial intelligence and true machine learning applications are still being tested, but the results so far have been extremely promising.
In a push to better identify and categorize images used in searches, Google’s Deepmind AI platform has reduced its image identification error rate to 5% – equivalent to that of human error. The potential applications for search are huge, but that’s not Google’s endgame. Over the last two years, Google researchers have been testing their AI systems using video games from the original Atari to help the computers learn to think for themselves and improvise solutions to common problems.
At Facebook, AI systems are using the company’s vast wealth of image data to help describe images to the blind.
Microsoft has developed (and demonstrated) a real-time language translation system in Skype to help bridge the language barrier during web conferences.
It’s clear that we’ve yet to scratch the surface on AI and machine learning systems, but the advances made and demonstrated to the public have shown remarkable progress in the still-burgeoning industry, leading the way for future innovations and advances for companies of every size.
What’s Available for Small Businesses
Big Data firms are enveloped in both a digital arms race for user adoption and against a ticking clock. As AI and machine learning moves ever forward, the useful aspects of human calculation and data-gathering becomes more and more obsolete. Concurrently, the value of knowledgeable data analysts who can provide insight into a company’s short and long-term predictive data capabilities will only grow as more organizations adopt sophisticated data analysis strategies.
Right now, however, the answers aren’t quite so clear. According to a recent report by IBM, only 60% of people in leadership roles have predictive analytics capabilities as part of their decision-making processes despite 3/4 of those leaders citing data analytics as a key source of future growth in their organizations.
In an interview with Techradar, Sean Owen, Director of Data Science at Cloudera, suggested that the exorbitant costs of AI and machine learning processes will result in small and medium-sized companies turning to their existing vendor relationships – Google, Amazon, IBM, etc. – for their AI and machine learning needs.
“SMEs are taking advantage of AI and ML (Machine Learning) processes through their existing vendor relationships. For example, using an expense solution like Concur allows an SME to take advantage of their Fraud Detection and Audit capabilities which leverage ML algorithms, or with rich relevance they can incorporate personalized interaction capabilities into their customer’s discovery and purchase experience. These “as-a-service” offerings are bringing the capabilities of AI to the SME where they can plug these capabilities into their business.”
Last year, Amazon launched a cloud-based service that allows companies to build predictive models using their existing data for a fraction of the cost of building their own infrastructure and systems. This provides a substantial opportunity for startups, small businesses, and even individual entrepreneurs to take advantage of huge data sets in building statistical analysis and predictive modeling and scale up as their organization grows.
But experts say first-time AI and machine learning initiatives at small companies shouldn’t begin as ambitious, sprawling efforts. Instead, augmenting the things small businesses do best (local-focused, flexibility, and friendly customer service) should be top priorities when building in a more comprehensive data analysis and predictive modeling strategy – but AI initiatives can only truly be successful if a company already has up-to-date and forward-thinking IT systems already in place.
What This Means Going Forward
Forrester, a leading research and analysis company, predicts that 16% of jobs in the U.S. will be lost over the next 10 years as a direct result of artificial intelligence and related technologies.
On the other hand, the firm believes that almost 14 million jobs will also be created during the same period of time in order to accommodate and compliment the burgeoning AI and machine learning industries.
Amid the flashy headlines and PR-filtered pull quotes, the undercurrent from all major players in the AI business is the same: don’t focus on replacing humans with AI systems; focus on improving human performance with AI. In stark contrast to the Terminator-centric AI/Skynet fearmongering that plagues the Internet, this goal is as positive and as optimistic as humanity can hope for.
Images labelled for reuse