Fawziah Selamat's TODAY article 'The atom calls to KL, Jakarta' (2007-01-06) made a rather Panglossian assertion: "...nuclear experts agree that an accident like Chernobyl's could not now take place in a new nuclear plant. Modern safety systems would make a total meltdown impossible. But there is always the niggling worry of human carelessness."

Which nuclear experts? According to advocates of Normal Accident Theory [NASA PDF], accidents in complex systems, such as nuclear power plants, are not just likely - regardless of safety systems, funding, training - but they are inevitable and thus 'normal'.

The influential landmark study of NAT by Charles Perrow, Normal Accidents: Living with High-Risk Technologies, shows how operator error exacerbated the Three Mile Island disaster but not because of human carelessness. The system was so complex and its interactions so tightly coupled that the instrumentation misled the operators into taking the wrong actions or the right ones only when it was too late. Perrow also reminds us that the accident at Chernobyl was caused by the unexpected reactions to the testing of a new, more modern and thus supposedly 'safer' safety system.

Newer and more modern technology often helps to make life easier, more productive and safer. But it can also work in reverse when its complexity overwhelms even technologically advanced US, Soviet and Japanese engineers. A more sophisticated understanding of industrial accidents, risk and complexity can help counter false assumptions that 'new technology is always better' and that we can eliminate risk that does not arise from human carelessness with any degree of certainty.