Earlier this year users of Link (the computer-based information system for lawyers) were confronted by an unusual message when they logged on to their screens.

Awaiting them was an apology and retraction, relating to comments made earlier about a charitable organisation and 'posted' to a bulletin board (ie made available to all users) on the Link system.

The comments were critical of the organisation and might have been defamatory in content.

In April, newspaper reports indicated that the ASDA supermarket group had settled by an out-of-court payment a libel action brought by a policeman, who claimed his reputation had been damaged by a message about him sent on the organisation's internal communication network.These two events illustrate the potential of electronic mail as a source of liability in defamation.

Such liability may impact not only on the authors of defamatory comments but also on their employers and the system operators who provide the technology which makes e-mail possible.

In the UK there are no decided cases but experience in the USA suggests that we will not have to wait long for claims to reach the courts.It should come as no surprise that messages sent by computer can be defamatory; telegraph and telex messages can be, and there is no reason in principle for computers to be seen as different.

Whether the defamation is libel or slander is more debatable, but material stored on diskette or in memory is permanent enough to make the former the better analysis.Many lawyers, however, might think that insuperable problems of proof stand in the way of any successful action based on electronic mail.

Messaging systems are used to carry all sorts of transient data, and defamatory gossip relating to third parties is unlikely to be saved by the sender or recipient.

Unless e-mail messages are deliberately saved or printed out, they are easily and apparently permanently removed from the screen by a key-stroke.

When removed in this way, the message does not appear to leave any trace behind.

But, in fact, it does and, to adapt a phrase, what you see is not what you get.

E-mail messages apparently removed from display are usually stored within the system, for an indeterminate time, and can be re-generated by anyone with the requisite technical knowledge and authority.

Sending e-mail is thus not like using the telephone.

If there is an analogy, it is with a telephone that has a recording device attached.Most casual users do not need to know about this.

But someone who considered him or herself defamed would have an interest in triggering the archival facilities built into any e-mail system.

Furthermore, there is no good reason why discovery should not be sought of such stored messages in appropriate cases, to facilitate proof of relevant facts in litigation.It is elementary that a potentially defamatory remark only becomes so when it is published to a third party.

Thus an e-mail message sent only to the person defamed and not (as in the Link or ASDA examples) made available to users of the system cannot give rise to any liability.

It is not 'publication' in the required sense for the addressee to pass on the message.

But would it be enough if the mess age were read by a system operator, perhaps in the course of his ordinary duties? Could it be said that to send a defamatory message by e-mail is equivalent to sending it by postcard? An author must accept that sending an openly accessible message is sufficient evidence of 'publication' (see Sadgrove v Hole [1901] 1 KB 1).

It may be that the courts would take the view that e-mail is inherently lacking in confidentiality as a medium, but the analogy with a postcard is not all that convincing.

While postcards will often be read by persons other than the addressee, the same cannot be said with e-mail users.

This being so, actual proof that a third party has read the message is probably required in all cases.The government's Defamation Bill, published this summer, sets out an extended defence of innocent dissemination which will go some way towards limiting the liability of non-negligent service providers for defamatory messages carried on their systems.But, for employers, including many law firms, practical solutions to the defamation problem must concentrate on alerting the users of computer systems to the dangers which the law of defamation presents.

This can be done in various ways.

Staff training programmes for computer users should draw attention to the risks involved.

Disciplinary rules ought to make it clear that computer facilities are not to be used to send inappropriate messages, and that sanctions may be imposed if they are.

It would do no harm to remind staff of the facilities which the system has to store and re-generate messages sent.

In the case of staff who fail to heed warnings given, dismissal may even be considered in extreme cases.

The law of unfair dismissal has established that it is likely to be fair to dismiss an employee who misuses a computer to obtain confidential information to which he is not entitled provided that due warning has been given (see Denco Ltd v Joinson [1991] IRLR 63).

There is no reason why a similarly tough line should not be taken with those who make improper use of the system after being told not to do so.Few employers would be content knowing that their staff were using office facilities for sending lies, gossip and scurrilous remarks to others within the organisation, far less to outsiders.

What might, however, have simply have been frowned on as time-wasting and anti-social conduct may have far more serious implications as defamation becomes a live issue.