This presentation identifies a number of potential sources or causes of animal welfare problems in Animal-Assisted Activities/Therapy (AAA/T) and assistance animal work. The goal is to focus attention on specific practices that may give rise to ethical concerns, and which ought therefore to be subjected to further scrutiny and study. The following preliminary recommendations are made:

 Those involved in preparing or using animals for service and therapy need to educate themselves regarding the particular social and behavioural needs of these animals, both to avoid the consequences of social and behavioural deprivation, as well as to permit animals a degree of control over the levels of social and environmental stimulation they receive.

 Animal-assisted intervention (AAI) practitioners need to understand that close physical contact with strangers may be inherently stressful for many animals, and recognise the signs of stress when they appear. Ideally, visitation and therapy sessions should be terminated before, rather than after, such symptoms are manifested.

 In residential programmes, one or more staff persons should be held primarily accountable for the care and welfare of any therapy animal, and for supervising all interactions with inmates/residents. No animal should be left unsupervised in a situation where its welfare might reasonably be considered at risk.

 Non-domestic species should not be used for AAA/T or assistance work except under exceptional circumstances (e.g., wildlife rehabilitation), and where appropriate care can be guaranteed.

 On the basis of current evidence, so-called 'dolphin-swim' programmes cannot be ethically justified.

 During the process of rearing and training assistance animals, transitions between successive handlers or owners should be carried out in such a way as to cause minimal distress due to the disruption of pre-existing social bonds.

 Efforts and resources should be dedicated to developing methods of accurately identifying and distributing suitable assistance animals from among those relinquished to animal shelters. These efforts should include research into appropriate behavioural screening methods.

 The present level of assistance dog 'failure' is ethically unacceptable and needs to be reduced. The 'industry' should be more aware of the problems inherent in the use of closed, purebred populations of service and assistance dogs. The potential benefits of out-crossing to other populations, and of crossbreeding, should be explored to reduce the prevalence of deleterious genetic diseases, as well as to improve infectious disease resistance.

 The 'industry' should give more attention to ensuring that assistance and service animals are adequately prepared during early development for the tasks and roles assigned to them as adults.

 Alternatives to the use of aversive conditioning in the training of assistance animals need to be investigated and developed wherever possible, particularly with respect to the training of so-called wheelchair dogs. If necessary, the 'industry' should consider discontinuing the use of animals for particular purposes, if alternatives to aversive conditioning cannot be found.

 More attention should be given to the design and construction of animal-friendly equipment and holding facilities for AAA/T and assistance animals.

 Continuing education programmes for animal practitioners and end-users should be available to ensure that animals are correctly handled, cared for and used.