The magic recipe for making AI possible? A sprinkle of cloud, intelligence and data
While AI is often hyped, it is starting to take off right now because the cloud, intelligence and data are coming together to make it possible, according to Pavandeep Kalra, director of data science, artificial intelligence and research at Microsoft, speaking Tuesday at The AI Summit in New York.
Take the cloud, for example, which is making advanced systems attainable to businesses of all sizes. "Up until now, computational power [was] hard," Kalra said. "If you wanted to actually spin up thousands of nodes on-premise on your cloud server, that's a very hard thing to do. So we now have pretty much infinite capacity on the cloud." The increased capacity allows the cloud to serve as an integration hub, helping break down data silos and make information readily accessible across teams.
While there is fear that intelligent systems will gain too much power, the true promise of AI will be to enhance human understanding. AI will work to increase understanding with the integration of diverse data sets and boost reasoning with the ability to tap into historical data, according to Kalra. Eventually, there will also be the potential for humans and machines to interact in a more natural way.
It is not uncommon for expectations of technology to far outpace reality. Take blockchain, for example. While the ledger technology shows a lot of promise, experts don't expect it to reach maturity until 2030.
In many ways, the same is true for AI. While intelligent and advanced systems are already used in production in businesses across sectors, technologists have a tendency to hype AI's capabilities, thinking first of the long-term promise it shows.
Key to managing expectations is understanding what is available now with AI and ML. The underlying technology is available to start launching AI, but companies are still migrating systems to the cloud and rely heavily on legacy architecture.
The vast majority of companies will not build intelligent systems in-house and will instead look to vendors to help implement AI capabilities. Amazon Web Services is one such vendor promoting advanced capabilities in an effort to draw customers onto its platform.
Moving past AI hype, businesses need to decide what is the use case for AI and establish potential measurable outcomes. Now that the technology is possible, it is important that organizations ready their systems and think of small-scale applications before attempting to transform the entire business with intelligent systems during a single implementation.
Follow Naomi Eide on Twitter