AI in a New Millenium: Obstacles & Opportunities

2 AI in a New Millenium: Obstacles & Opportunities 2 2.1 AI as both Science and Engineering . . . . . . . . . . . . . . . . . . . . 2 2.2 Difficulties and how to address them. . . . . . . . . . . . . . . . . . . . . 4 2.3 Institutional and financial obstacles . . . . . . . . . . . . . . . . . . . . . 4 2.4 Failing to see problems: ontological blindness . . . . . . . . . . . . . . . 5 2.4.1 What are the functions of vision? . . . . . . . . . . . . . . . . . 5 2.4.2 Seeing without recognising objects . . . . . . . . . . . . . . . . 6 2.4.3 Is language for communication? . . . . . . . . . . . . . . . . . . 8 2.4.4 Varieties of complexity: ‘Scaling up’ and ‘scaling out’ . . . . . . 9 2.5 Are humans unique? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2.5.1 Altricial and precocial skills in animals and robots . . . . . . . . 12 2.5.2 Meta-semantic competence . . . . . . . . . . . . . . . . . . . . . 13 2.6 Using detailed scenarios to sharpen vision . . . . . . . . . . . . . . . . . 14 2.7 Resolving fruitless disputes by methodological ‘lifting’ . . . . . . . . . . 15 2.7.1 Analyse before you choose . . . . . . . . . . . . . . . . . . . . . 15 2.7.2 The need to survey spaces of possibilities . . . . . . . . . . . . . 16 2.7.3 Towards an ontology for types of architectures . . . . . . . . . . 17 2.8 Assessing scientific progress . . . . . . . . . . . . . . . . . . . . . . . . 18 2.8.1 Scenario-based backward chaining research . . . . . . . . . . . . 19 2.8.2 An example collection of scenarios . . . . . . . . . . . . . . . . 19 2.8.3 Assessing (measuring?) progress . . . . . . . . . . . . . . . . . . 20 2.8.4 Replacing rivalry with collaboration . . . . . . . . . . . . . . . . 21