Innovation in Machine Intelligence May Be Unstoppable
The Leading Edge Forum contends that we are fast approaching a point where machine (artificial) intelligence will develop rapidly and impact every industry because crucial factors have aligned.
- By Lindsay Stares
- June 24, 2016
The Leading Edge Forum, a global research and thought-leadership community, released a position paper outlining its belief that what it calls machine intelligence (aka artificial intelligence) is nearing a tipping point. They claim that due to the alignment of specific factors, the conditions are now right for rapid innovation in machine learning.
Formula for Innovation
The formula they propose: big data + new software and hardware + cloud computing = rapid innovation in machine intelligence (MI). It's certainly plausible.
Huge amounts of unstructured data are being created every day, and this content provides the raw material needed to develop and test machine intelligence. Preliminary big data applications for MI include enabling computers to recognize images or identify emotion in text.
New designs for multi-layered deep learning designs allow systems to break down complicated tasks and handle them in greater detail. New hardware designs provide the processing power needed for these advanced applications.
The rise in cloud computing is providing the storage necessary for these computation-intensive tasks as well as the drive to develop this technology. The advantages of the cloud -- rapid deployment, real-time data, enormous scalability, etc. -- are an ideal match for the potential of advanced MI, and enterprises working with large amounts of data in the cloud need the power and scope of MI applications.
Impact on the Enterprise
LEF sees huge opportunities for MI at both start-ups and established tech giants. Effective machine intelligence already delivers search results, translates languages, and identifies your friends in your Facebook photos. If innovations create smarter software that can more accurately analyze and forecast, there are potential applications in every industry.
The researchers at LEF recommend that enterprises prepare for machine intelligence by thinking about it now. Start by considering what applications are relevant to your field or what parts of your business could be run by algorithms. They also recommend evaluating your data sets for MI potential, identifying employees who will lead in MI, and keeping an eye on developments in the field.
The World on MI
It seems probable that machine intelligence will improve at least somewhat as predicted, given the enabling technologies and the economic incentives. However, what that will look like in everyday life is still an open question.
The LEF paper acknowledges that once a new technology becomes common, people are very good at accepting it as just another service. Even some of the deep learning applications already on the market, personal assistants such as Siri and Cortana, are beginning to seem like just another thing that phones and computers do rather than ground-breaking technology.
This tendency to adjust to new technology is opposed by the tendency to mistrust new technology. We may come to a point when society will balk at handing over certain functions to intelligent machines. Where each person and each culture draw that line may vary; will you accept self-driving cars? Algorithms grading school papers? A news report written by a program?
LEF believes that the potential benefits of advanced machine intelligence will outweigh the negatives, but we can't fully predict the impact of MI. For more details, see the complete paper on leadingedgeforum.com.
Lindsay Stares is a production editor at TDWI. You can contact her here.