2017 Ethics of Predictive Cyber Security
March 16, 2017 by The GTech Marketing Team
Predictive cyber security refers to the application of stochastic models to the cognitive analytics of safeguarding computerized information. Stochastic programming bases decisions on probability distributions, probabilistic decisions, and the optimization of random variables, but artificial intelligence dominated Dell's 2017 RSA cyber security conference. Artificial intelligence refers to the human aspects of intelligence. Artificial intelligence requires human intuition and deep learning to make decisions.
In 1997, a supercomputer processed 200 million chess positions per second. Improvements in computer architecture, programming, and engineering enabled computers to "learn." In 2007, a supercomputer processed game moves continuously "learning" strategies and tactical contingencies. Like artificial intelligence, machine intelligence approximates human learning through trial and error, guessing, and modifying moves as play advances to achieve the best possible outcome.
Although Dell regards machine intelligence a new cognitive era in data processing, automation, and robotics, the use of machine intelligence in cyber security solutions poses ethical issues. Cisco Security’s David Ulevitch prefers safe guards on machine intelligence in its 2017 stage of development. Automation results in loss of human jobs, but it's cost-effective compared to hiring 500 thousand cyber security techs nationwide.
Dr. Evan Selinger, professor at the Rochester Institute of Technology, stressed the lack of actual data to predict the outcome of machine-based cyber security decisions. Machine learning also results from the process of identifying patterns in massive amounts of data. Digital technology simulates neural networks in the brain and processes information based on what it learned without direct computer programs directing its actions. Another ethical consideration is whether users have the right to know cyber security algorithms or they are proprietary information, intellectual property of the originator. David Auerbach pointed out that acquiring an algorithm is not the same as understanding an algorithm.