5.2.2 AI, Machine Learning & Robotics (B05)

a machine capable of carrying out a complex series of tasks automatically robotics is the design, construction, operation and use of robots

the ability of a digital computer or computer-controlled robot to peform tasks commonly associated with intelligent beings (those that can adapt to changing circumstances)

ability of machines to learn on their own without being programmed - from data patterns - trained on data (not ‘fed’) symbolic learning is when humans feeds AI with explicit rules/logic/symbols that they can understand

able to make decisions without human involvement or guidance

AI decisions may be incorrect because they are biased against certain groups of people - introduced in biases in training data - may be unrepresentative of reality or reflect existing undesirable prejudices eg Amazon’ recruitment tool was trained on historical hiring decisions which favoured men over women - computer then repeated human bias

basically the trolley problem but for autonomous vehicles: if an autonomous vehicle swerves and hits 3 people to save 5, is that moral? and if it isnt, who is accountable? the non-driving owner, the manufacturer, the algorithm developer? should each country makes its own rules in terms of how to handle these problems? or should there be a worldwide body? there is presently not agreement (apart from that the manufacturers of the vehicles should not make these rules)

search for and engage targets on their own without direct human control eg: drones, tanks, submarines, ‘killer robots’ 30+ countries (Austria, Brazil, Canada) have asked to ban these but proponents say: - they will make wars more ‘humane’ - they will be more precise and efficient - they wont fall prey to human emotions and will minimize deaths of soldiers • facial recognition may be inaccurate • may kill innocent civilians to achieve goals • vulnerable to interference and hacking - then who responsible?

arent many laws formulated to control development or give definitions of responsibility at present there is a rule that say a person on behalf of whom a program was creates must, ulitmately, be liable for any action generated by the machine - but it is unclear whether this will be applied to driverless cars and autonomous weapons should make laws on who is responsible in these cases. should also make laws controlling development - how will it affect employment, economoic inequality / wealth gap.

throughout history introduction of technology has led to societal problems - unemploymet, starvation, rioting AI can automate repetitive and even more complex tasks - these jobs will become obsolete