Is digital evidence as reliable as we think?

I have been following this case with interest for many years. In 1999, the UK Post Office introduced a new computer system called Horizon to record transactions, accounting and stocktaking.

Soon after sub-postmasters complained that the software was buggy, leaving them with shortfalls of many thousands of pounds. At the time, the Post Office refused to accept that Horizon could be wrong and threatened many sub-postmasters with legal action. Some remortgaged their homes and took out other loans to avoid legal action but between 2000 and 2014 the Post Office prosecuted 736 people for false accounting and theft based on evidence from Horizon.

The lives of many people we were ruined; sent to prison, financially ruined and shunned by communities and their families. Finally, as reported by the BBC below, after over twenty years, Horizon has been found to contain bugs and defects that caused these deficits.

I admire the sub-postmasters who maintained their innocence after all this time, and kept the energy to keep fighting. At times it must have felt like they were against an impossible brick wall and I really feel for the people who sadly died before seeing their names cleared.

No doubt for years to come people will be asking - what went wrong and how can we make sure something like this never happens again? We already acknowledge that computer-based evidence is very fragile and, in some cases, present for only a short period of time. In the UK the Association of Chief Police Officers (Now the National Police Chiefs' Council) Good Practice Guide for Computer-Based Electronic Evidence is designed to ensure that the evidence is captured accurately, completely and presented without alteration.

However, I believe that we are too trusting of computer-based evidence. The capability to perform calculations that were undreamed of only a few decades ago, combined with the fact that they faithfully execute their instructions I feel has contributed to this elevation. Forgetting that the input data or instructions may be wrong, either accidentally (bugs) or deliberately (hackers). 

In conclusion, I think the courts, and the public, need to be more sceptical of computer-based evidence and require corroboration to a similar standard as other types of evidence, after all faulty data correctly captured, preserved and presented is still faulty data.