>From The Irish Times,
Friday, January 28, 2000
Study offers glimpse
of robotic future
---- By Madeleine Lyons
WIRED ON FRIDAY: There are rare moments when developments in the world of information technology (IT) come under the spotlight and the thin line between reality and fiction becomes blurred.
A new study "Morals for Robots and Cyborgs" reads like a modern day version of Franken- stein. It offers a rare glimpse of the potential monster man may have inadvertently created in the quest for greater convenience and technical efficiency.
The study, commissioned by the Bull Cara group, examines some of the ethical and public policy issues surrounding the use of advanced computers, especially robotics, digital agents and neuro-computer links.
The primary aim of the report is to stimulate early debate on areas of potential public concern that have been created by technology, and suggest possible policy guidelines.
According to Mr David Little, managing director of Bull Cara Group: "It is vital for the computer industry to open a dialogue with the public, government, industry and academia if we wish to avoid the controversy that dogged other sectors like genetically modified (GM) foods, cloning, human fertilisation and trans-species organ donation."
The central plank of the report is based on the writing of the rather unusually named Perri 6, a senior research fellow at the University of Strathclyde.
Despite the bizarre moniker he took for himself more than 20 years ago, Perri 6 is a respected academic who rose to prominence through his involvement as a director of Demos, the UKbased think-tank, which has been closely identified with supporters of New Labour.
His study investigates the challenges for ethics and public policy raised by technical advances. He outlines 15 principles of "techno-ethics" which range from the need to monitor the economic effects of deploying artificial intelligence, to ensuring we can manage and compensate the changes in demand for human labour that may result from industrial thinking machines.
He points out that concerns about new technologies, often prompted by the ideas behind science fiction movies, may be misplaced, and the likelihood of a world where intelligent machines like HAL, RoboCop or Marvin the Paranoid Android are going to seize power is somewhat overstated.
Instead he considers the potential economic impact of autonomous artificially intelligent systems, the need to ensure privacy, health and safety and the need to educate the wider community about advances in general.
"For most of history, machines have been tools. Now they have the capacity to make decisions, what kind of legal and ethical rules do we need to govern them? A neural Net can provide some answers to medical problems as efficiently as a doctor, but when they get it wrong they tend to kill the patient. There are fundamental questions about liability, as well as privacy and desirability - is a machine used by a criminal more like a gun or a car?" asks Perri 6.
The academic investigation spans a number of areas from existing technologies - intelligent software for example which refines its decisions and behaviour through learning - to emerging technologies which predate scenarios where wars might be conducted by robots on our behalf.
One of the more practical conclusions makes the rather obvious point that with technology replacing whole areas of formerly human responsibility, there are implications for human labour. Perri 6 advises that politicians should be prepared to undertake compensating measures to stimulate employment opportunities for those who may be displaced by substitution. With traditional retailing and supply chain management in a current state of overhaul the advice is timely.
The study touches on other, more notional, concerns including the moral conduct of war by machines and regulatory issues for government. Perri 6 is adamant there is huge potential for misuse of knowledge-based systems.
"Robots and digital agents could be used in the course of crime. The largest category of investment in digital agents is in the military where they could be used to replace soldiers. On one hand they are less likely to desert or be captured, on the other, what happens if a machine turns in the battlefield. Accountability and responsibility also need to be designed in," Perri 6 says.
There are also issues of privacy in cases where intelligent software is deployed to learn over time about an individual's behaviour online, including transactions and Internet tastes and preferences.
This information once accurately identified is extremely commercially valuable. Perri 6 calls for principles of software design and disclosure which will protect the consumer from commercial predators.
In conjunction with the paper a survey of public opinion was conducted by Taylor Nelson Sofres, from 1,001 face-to-face interviews which took place in the UK last October.
The survey's findings indicate an alarming disparity between male and female views on technology's progress.
For example 15 per cent more women than men were concerned that machines had already taken too many jobs (70 per cent), and 12 per cent more men than women believed technological advances to be generally a positive thing (72 per cent).
Techno-fear featured strongly in the survey, with more than half the respondents agreeing we are in danger of entering a robot state where everything is automated, and 81 per cent disagreeing with the principle of embedding chips into humans.
However, the inevitability of technological progress was largely agreed on, with more than 70 per cent of respondents unconcerned about the spread of technology, believing artificial intelligence to be generally a good thing.
One of the most interesting conclusions of the survey, was the finding that nearly half (48 per cent) said they would trust the government least to monitor technological advances. Academia scored high here, coming in as the most trusted monitoring body for 25 per cent of respondents, while the computer industry would only be trusted by 14 per cent if it were to assume this role.
Both the study and survey offer a refreshing glimpse of what lies ahead for information technologies, and point to some of the interesting debates that need to be thrashed out. Bull Cara's involvement in the project indicates a renewed interest by the European IT giant in industry progress. The company needs to make up some ground it lost since it came to dominate the European market in the early nineties, particularly in smart card and mainframe systems.
Perri 6's paper "Morals for Robots and Cyborgs" and the ac- companying research can be viewed on Bull's website: www.bull.co.uk.
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:02:52 MDT