More data, more problems
As tech generates more data, smarter tools help inform better decisions
It was an algorithm gone wrong.
The colorful flyer called out, "Congratulations! On Your New Baby!” Yet the couple who got it were retirees. Software mistook them for parents-to-be because they had bought baby clothes for a new grandchild at a big-box retailer.
A human analyst might have picked up on other clues and avoided that mistake. The lesson: Technology still needs the human touch to work right, even though the day is coming when it may be able to come up with near-human judgment calls.
“Advanced, machine-learning algorithms can help businesses make sense of all the data being captured quicker and less expensive than humans,” said Mark Bigham, chief innovation officer for Raytheon's Intelligence, Information and Services business. “Analytics and artificial intelligence are right on the edge of delivering near-human accuracy and insight.”
Raytheon is developing and deploying analytics tools to meet specific needs for both government agencies and commercial enterprises. Technology like the company's Intersect analytics software suite mix high-performance computing, big data storage and sophisticated algorithms to achieve a reasonable and responsible level of automation.
But data analysis has its limits. Machines can root through a deluge of information, but their best function is to help people verify findings and make the right decision.
The Human Element
Analysts today are collecting more and more data from a vast, growing array of sources, including retail transactions, online exchanges and the Internet of Things.
“Commercial companies and government agencies use, capture and store massive amounts of data, including our personal data,” said Jane Chappell, Raytheon IIS vice president of Global Intelligence Solutions.
As technology continues to push toward a day when machines will make decisions independently, humans who build and operate them will have to evolve as well.
“We’re currently living at the beginning of the fourth industrial revolution, where massive amounts of data are being captured and stored on low-cost devices with great computing power,” Bigham said.
One major challenge is protection of the data being captured. Storage systems must be cyber-hardened to ensure security, including potential for damage from insider threats.
Protecting the Data
With trillions of bytes of data being gathered through open-source collection methods, social media sharing or customer loyalty programs, the challenge is to identify and safeguard that information even as it is being captured, according to Guy Swope, Raytheon Analytics Capability Center leader.
“Protecting data can be expensive, but it’s important that all the data captured is protected,” Swope said. “Companies and governments that fail to protect data sufficiently could be unwillingly inviting hackers to take important data.”
“While the data can be used to identify waste, fraud, and abuse, a criminal actor could use that same data to buy a new boat in your name,” Chappell said. “It’s important that all organizations employ a cyber secure analytics solution, capable of both assisting in decisions and protecting information at the highest standard."