A session at last week's International Bar Association conference in San Francisco raised the intriguing possibility of lawyers receiving calls for representation by sentient computers in a few years.

An innovative mock trial on 'biocyberethics' started with the premise that, by the year 2020, computers will have achieved human-level intelligence.

It then looked at what lawyers should do if such a computer sought legal advice to prevent others turning off its power (equivalent to cessation of life support), changing its programming without consent or causing it pain via technical experiments.

A lawyer for the computer argued that once it becomes intelligent enough to show initiative, it deserves rights.

For the company, it was argued that a computer can only simulate feelings.

The audience, acting as the jury, voted in favour of an injunction sought by the computer to stop its owner pulling the plug and 'killing' it.

However, in the tradition of US legal dramas, acting judge Joseph McMenamin of McGuire Woods overruled them and said he would reject the application for an injunction, but prevent the company turning off the computer pending an appeal.