Noel E. Sharkey

Granny and the robots: ethical issues in robot care for the elderly

The growing proportion of elderly people in society, together with recent advances in robotics, makes the use of robots in elder care increasingly likely. We outline developments in the areas of robot applications for assisting the elderly and their carers, for monitoring their health and safety, and for providing them with companionship. Despite the possible benefits, we raise and discuss six main ethical concerns associated with: (1) the potential reduction in the amount of human contact; (2) an increase in the feelings of objectification and loss of control; (3) a loss of privacy; (4) a loss of personal liberty; (5) deception and infantilisation; (6) the circumstances in which elderly people should be allowed to control robots. We conclude by balancing the care benefits against the ethical costs. If introduced with foresight and careful guidelines, robots and robotic technology could improve the lives of the elderly, reducing their dependence, and creating more opportunities for social interaction

Combining diverse neural nets

An appropriate use of neural computing techniques is to apply them to problems such as condition monitoring, fault diagnosis, control and sensing, where conventional solutions can be hard to obtain. However, when neural computing techniques are used, it is important that they are employed so as to maximise their performance, and improve their reliability. Their performance is typically assessed in terms of their ability to generalise to a previously unseen test set, although unless the training set is very carefully chosen, 100p accuracy is rarely achieved. Improved performance can result when sets of neural nets are combined in ensembles and ensembles can be viewed as an example of the reliability through redundancy approach that is recommended for conventional software and hardware in safety-critical or safety-related applications. Although there has been recent interest in the use of neural net ensembles, such techniques have yet to be applied to the tasks of condition monitoring and fault diagnosis. In this paper, we focus on the benefits of techniques which promote diversity amongst the members of an ensemble, such that there is a minimum number of coincident failures. The concept of ensemble diversity is considered in some detail, and a hierarchy of four levels of diversity is presented. This hierarchy is then used in the description of the application of ensemble-based techniques to the case study of fault diagnosis of a diesel engine.

Word recognition in a functional context: the use of scripts in reading

Abstract The interpretation of contextual effects on the recognition of words has generally been isolated from any high-level functional theory of context. The possibility of using one such theory, that associated with the script construct, was investigated in four lexical decision experiments. A clear-cut pattern of facilitation for associated script words was demonstrated in Experiment 1 and replicated in Experiments 2 and 3. In Experiments 2 and 3 this facilitation was shown to be robust with respect to intervening materials up to three sentences long. The results from Experiment 3 also suggest that the influence of script contexts does not simply decay over time since last mention; rather it is deactivated by control cues in a text. Further, the findings from Experiment 4 imply that under certain circumstances at least two scripts may be active at the same time. It is argued that script contexts could be useful for incorporating word recognition research with research on text comprehension. An associative network account of the findings is given.

Weak contextual constraints in text and word priming

Abstract The effects of text priming and word priming have often been confounded in the literature. Four lexical decision experiments are reported here that clarify this distinction. In the first two, priming effects are observed for words related to a prior text even though they immediately follow an unrelated word. Two further experiments dismiss the possibility that this effect is attributable to sustained priming, stemming from associatively related words occurring earlier in the text. Overall these findings run counter to predictions from Kintschs (1988, Psychological Review, 95 , 163–182) construction-integration model and favor the lexical distance model Sharkey, 1989a , Sharkey, 1990a . However, an absence of associative priming for prime target pairs embedded in text rather than scrambled sentences (Experiments 3 and 4) is a difficult finding to align with any current model of word recognition.

The Ethical Frontiers of Robotics

The use of robots to care for the young and the old, and as autonomous agents on the battlefield, raises ethical issues.

Saying ‘No!’ to Lethal Autonomous Targeting

Abstract Plans to automate killing by using robots armed with lethal weapons have been a prominent feature of most US military forces’ roadmaps since 2004. The idea is to have a staged move from ‘man-in-the-loop’ to ‘man-on-the-loop’ to full autonomy. While this may result in considerable military advantages, the policy raises ethical concerns with regard to potential breaches of International Humanitarian Law, including the Principle of Distinction and the Principle of Proportionality. Current applications of remote piloted robot planes or drones offer lessons about how automated weapons platforms could be misused by extending the range of legally questionable, targeted killings by security and intelligence forces. Moreover, the alleged moral disengagement by remote pilots will only be exacerbated by the use of autonomous robots. Leaders in the international community need to address the difficult legal and moral issues now, before the current mass proliferation of development reaches fruition.

What is the point of integration? The loci of knowledge-based facilitation in sentence processing ☆

Abstract When people read a text they must construct a representation of its essential ideas and integrate them. We investigated how world knowledge facilitates this process, and examined the loci of integration during sentence processing. In four experiments, we manipulated the relatedness of sentence-internal targets to preceding knowledge-based texts. In the first three experiments, no differences were observed in any sentence position except the sentence-final position. These results indicate that the integration of entire sentences with prior text is facilitated by active knowledge structures. Explanations in terms of associative priming, transitional probability, and the activation of anaphoric referents were ruled out. In Experiment 4 a difference was observed in the initial position between clearly unrelated control targets and knowledge-based targets. Two alternative accounts of the data are given: a sophisticated buffer model with a contextual checking mechanism, and an activation window model which is equivalent to a buffer with contextually sensitive access.

Connectionist representation techniques

Connectionist natural language processing research has been in the literature for less than a decade and yet it is already claimed that it has established novel styles of representation. This article presents a survey of some of the main representational techniques employed in connectionist research on natural language processing and assesses claims as to their novelty value, i.e. whether or not they add anything new to Classical representation schemes. The main aims are (i) to introduce readers (particularly AI researchers and computational linguists) to the nuts and bolts of the different styles of connectionist representations and (ii) to lay out the direction of research on the new uniquely connectionist representations. These latter representations hold a great deal of promise for the beginning of a new theory of Artificial Intelligence (AI).1

Children, the Elderly, and Interactive Robots

The idea of developing robot companions and caregivers for the elderly is taking hold. Elderly people are often lonely and in need of companionship and social contact. Some hold that a robot could be a friend substitute and, at the same time, reassure absent families about the well-being of their elderly relative by monitoring and reporting on their health. Alzheimers disease leaves many elderly confused so that they need help with routine activities and someone to answer their questions. It has been suggested that a robot could fulfill this role. Young children need constant care and supervision, but busy parents do not always have the time to provide it.In this article, we have probed the ethics of designing robots that promote the illusion of being able to form meaningful relationships with humans.

Cassandra or False Prophet of Doom: AI Robots and War

The US military currently deploys more than 4,000 ground robots in Iraq. Most are for bomb disposal but some are armed. In addition to the robots, semi-autonomous unmanned air vehicles come equipped with hellfire missiles. For now, all these systems employ a human-in-the-loop for the application of lethal force. But this is set to change. In the near- to mid-term future, the military might allow autonomous unmanned systems to make their own lethality decisions. This article probes the published US military plans and raises questions about the application of AI to discriminate between innocents and combatants in modern warfare. It points to the main ethical issues in terms of the laws of war and discusses the responsibilities of AI researchers embarking on military projects.