Article - Issue 13, August 2002
The pursuit of trustworthy information systems in the new millennium
Caspar Bowden, Jon Crowcroft and Roger Needham
Today and for the foreseeable future, new public policy issues are brought up by advances in communication and information technology. The Internet is much more than an apparatus for moving bits of information around the world – it generates whole new patterns of behaviour and whole new possibilities for everyone. ‘Everyone’ means what it says – the terrorist as well as the educator, the law abider as well as the criminal. Presented with a kaleidoscope of good and bad, our governmental, legal and regulatory infrastructures have to decide what to do.
In a democracy, such decisions should be made on the basis of informed debate between those responsible for these infrastructures and the public. Unfortunately, a number of the issues are quite technical. Actions that look obviously desirable may prove to be technically very difficult or expensive, or may turn out to inhibit conduct that is valued or enhance conduct that is deplored. There is no common currency in the weighing of enhanced law enforcement against loss of personal privacy. Legislation that was perfectly sensible in a former environment can turn out to be unsuitable now. New procedures for electronic commerce, for example, while aiming to make transactions quicker and easier, can produce an undesired and unmentioned redistribution of risk between participants.
Recent events, both creative, such as the successful cloning of higher species and the near completion of the human genome project, and destructive, such as the BSE crisis or attacks of 11 September, have put a great deal of information stress on the population. When experiencing such ‘future shocks’, it is common to react as if there are no relevant expertise or skills to bear and to make up policy ‘on the fly’.
In information systems, there is in fact a large body of expertise and a collection of skilled individuals and organisations who are able to draw upon previous situations and make sense of alternative reactions in a calm and considered way.
The way that government agencies, businesses, private organisations and individuals interact through information systems is evolving very quickly. The rapid introduction of databases describing us and our property, our history and our future (e.g. insurance and online transaction systems between organisations and individuals) has been faster than any rate of uptake of technology in the past. Compare the rate of change of the Internet and the Web with the timescales for the migration from early printing press to paperback books, or the millennia associated with the transition from the use of banknotes to the introduction of debit cards. Compare the time it took from its first use for the PC to become all-pervasive with the rate of transition from the use of papyrus to paper.
Modern manufacturing and communication systems not only enhance the pace of change, but also the complexity of organisational relationships. It is expected that cultural and business relationships will now be global. Many global companies now do no manufacturing themselves, but outsource the process in a number of places. Because of cultural, political, economic and legal differences, this leads to exponential growth in the complexity of the information associated with such relationships. There are no longer any real simple hierarchical chains of government and trust.
Privacy is a major issue. Correlation of different sets of personal data can seriously erode personal privacy. All too often we fail to define the use of data accurately. The Data Protection Act goes some way towards requiring the keeper of personal data to define use. However, there are exceptions and, even when not, there is little attempt to ask users to think about risk, cost, conflicts and roles of information users.
We do not ask people for a clear exposition of their expectations, for example, for reliability of information and change control. We do not have adequate tools to describe the behaviour of information users, nor for managing the acceptance of information use amongst subjects. Neither message secrecy nor deidentification of records is sufficient for privacy. Ill-informed people often see encryption as a panacea in this area, which it emphatically is not.
The public understanding of risk in security systems must include the notion also that there are no ‘zero defect’ approaches. The policy must make clear what are the priorities (privacy, safety, etc.) and the engineering of the systems must make transparent that the policy has been captured in the operational system.
‘The price of freedom is eternal vigilance.’ The price of information systems security (privacy, integrity, reliability, trustworthiness, affordability, usability, etc.) lies in the education and proper resourcing of the overall system. It is not simply a naïve combination of hardware with underpaid, undertrained, overloaded staff.
Most studies show that the vast majority of security system failures are caused by people – either accidentally, deliberately, or in complex combinations of unintentional and intentional actions.
Caspar Bowden, Jon Crowcroft and Roger Needham
The authors are members of the Foundation for Information Policy Research. FIPR (www.fipr.org) seeks to enhance the quality of debate by producing clear statements on the issues at stake and reasoned proposals for policy improvements.