The Causation Issue in Computer Security Breach Cases
Cherry, Michael, Imwinkelried, Edward, Judicature
The growing number of computer security breaches has led to a proliferation of lawsuits, including class actions, against the companies that have experienced breaches - and consequently exposed their customers and clients to the risk of identity theft. One of the key issues in such litigation is causation; What caused the breach, and could the company have prevented it?
The truth is that the system itself - our current approach to cyber security - is badly flawed. The large number of recent, large scale breaches of computer security - including Hannaford Farms, Heartland, and Countrywide - is stunning. Despite all die highly publicized efforts to improve cyber security, the situation seemingly continues to worsen.
Early computers used a limited set of commands that unintentionally minimized the possibility of hacking. The personal computer, which developed decades later, chose a very different approach. The PC used a very responsive interface that allowed suppliers and others to dynamically add new equipment and procedures through die use of simple instructions. Understandably, the vast majority of public and private sector organizations opted for PCs. The obvious upside was that PCs were less expensive to maintain because it was easier for support personnel to master them. However, the combination of the lack of rigorous authentication and the ease of adding new procedures had a significant downside: It made it virtually impossible to prevent hacking.
As the internet evolved, hacking became more widespread. Organizations attempted to block hackers by resorting to add-ons such as virus scanners, "strong" passwords, and network testing. Ever since suppliers have been on a treadmill, constantly providing users with updates for new add-ons. This reliance on add-ons has failed. It is easier and faster for a hacker to write a new virus than it is for the virus industry to identify an existing one and devise an add-on to counter it.
To prevent hacking, we could: (1) rewrite Windows and UNIX to make them less responsive, or (2) convert to a ncw operating system that is less responsive and conse- quently less vulnerable. Both efforts will require significant re-engineering and conversion. But in the meantime, whichever option we choose, we must immediately ensure that systems have uhe capacity to automatically recognize a breach in progress and the related capability to automatically lock down all sensitive information.
Consider the Countrywide incident. At Countrywide, an employee, a mortgage loan evaluator, brought home more than 5,000 difieren t mortgage applications for "review" every weekend for almost two years. Il would have been amazing if he had managed to thoroughly review even 50 files during a single weekend. Yet, he continued bringing home these massive quantities of sensitive information for an extended period of lime until his partner in crime, the buyer of the mortgage applications, was finally caught. By any rational business standard, it should have been obvious that large quantities of sensitive information were being inappropriately retrieved. But Countrywide's security system did not catch the mortgage loan evaluator because the system did not incorporate any specific retrieval limitations based on business standards.
In the instances of Heartland and Hannaford, large quantities of sensitive information were electronically sent to a third party. The existing computer security systems permitted the transmission of huge amounts of sensitive data to outsiders without sounding any alarm even though it should have been obvious to any objective observer that the massive transmissions deviated from the companies' normal business process.
Today's information protection procedures rely almost entirely on external measures such as tests for viruses and malware, passwords, and the use of white-hat hackers to identify network weaknesses. None of these measures factors in an understanding of the normal operations of the business-at-large. …