An Ontology for Cognitive Mimetics

Saariluoma, Pertti; Wahlström, Mikael; Kujala, Tuomo; Puustinen, Ronny; Karvonen, Antero; Happonen, Kai

doi:10.1109/csci46756.2018.00228

Cited by 4 publications

(12 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the trolley problem and its variations show, actual human ethical judgement is a complex affair which integrates many kinds of information processes and contents together against open-ended problems with many possible solutions. An AI system built around a single variable as the target of an ethical evaluation function may work for many cases but fail (from an ethical standpoint) for others because it has not mimicked [2][3][4][5][6] the actual information process in humans. For example, human beings judge based on emotions, or concepts like allowing vs. doing harm [20].…”

Section: Ai Ethics As a Design Problemmentioning

confidence: 99%

“…Thus, accessing by empirical means (without artificial limitations) how humans restructure problems [21] can give crucial hints on how to build similar abilities into AI systems, and discover usable patterns for moral problem solving. The empirical route sketched in cognitive mimetics [2][3][4][5][6] provides implementation cues for both general abilities and specific contents and patterns [see 6]. The whole point of the trolley problem is that within its limitations there is no right answer, and in such unfortunate circumstances it is difficult to see how either humans or machines should be forced to consider ethical questions, as both making the choice and not-making it result in an unethical action in one sense or another.…”

Section: Ai Ethics As a Design Problemmentioning

confidence: 99%

“…Cognitive Mimetics is an idea for a design method for intelligent systems introduced by Kujala and Saariluoma and elaborated over a series of papers [2][3][4][5][6]. Mimetic design means using a source in the natural or artificial worlds as an inspiration for technological solutions.…”

Section: Cognitive Mimetics For Ethicsmentioning

confidence: 99%

“…However, in creating intelligent technologies designers can use existing organizational and individual information processes as the source of ideas. Designing intelligent systems by utilizing existing human information processes as the source of solutions we have termed 'cognitive mimetics' [2][3][4][5][6]. Cognitive mimetics differs from typical and established biomimetics as it has different source of mimicking: human shared and individual cognitive processes, as well as the mental contents, representations, and constraints that establish the boundaries and forms it takes.…”

Section: Cognitive Mimetics For Ethicsmentioning

confidence: 99%

“…The second is where to look for the basis for ethical reasoning -where should "ethics modules" derive their contents and their basis for reasoning? This article offers a conceptual discussion on these issues and then lays out how cognitive mimetics [2][3][4][5][6] can be used in the context of AI ethics.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Cognitive Mimetics for AI Ethics: Tacit Knowledge, Action Ontologies and Problem Restructuring

Karvonen

2020

Culture and Computing

Self Cite

View full text Add to dashboard Cite

Ethics and ethical information processing are an important problem for AI development. It is important for self-evident reasons, but also challenging in its' implications and should be welcomed by designers and developers as an interesting technical challenge. This article explores AI ethics as a design problem and lays out how cognitive mimetics could be used a method for its design. AI ethics is conceptualized as a problem of implementation on the one hand, and as a problem of ethical contents on the other. From the viewpoint of human information processing, ethics becomes a special case of ethical information processing -one that has deep implications in terms of AI abilities and information contents. Here we focus on ethical information processing as a property of the system (rather as a general constraint on it). We explore three specific concepts relevant for cognitive mimetics from the perspective of ethics: tacit knowledge, ontologies, and problem restructuring. We close with a general discussion on the difference between abilities and mental contents noted as relevant in previous articles on cognitive mimetics and reiterate its importance in this context as well.

show abstract

Section: Ai Ethics As a Design Problemmentioning

confidence: 99%

Section: Ai Ethics As a Design Problemmentioning

confidence: 99%

Section: Cognitive Mimetics For Ethicsmentioning

confidence: 99%

Section: Cognitive Mimetics For Ethicsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Cognitive Mimetics for AI Ethics: Tacit Knowledge, Action Ontologies and Problem Restructuring

Karvonen

2020

Culture and Computing

Self Cite

View full text Add to dashboard Cite

show abstract

Hume’s Guillotine in Designing Ethically Intelligent Technologies

Saariluoma

2020

Advances in Intelligent Systems and Computing

Self Cite

View full text Add to dashboard Cite

Intelligent machines can follow ethical rules in their behaviour. However, it is less clear whether intelligent systems can also create new ethical principles. The former position can be called weak ethical AI and the latter strong ethical AI. Hume's guillotine which claims that one cannot derive values from facts appears to be a fundamental obstacle to strong ethical AI. The analysis of human ethical information processes provides clarity to the possibility of strong ethical AI. Human ethical information processing begins with positive of negative emotions associated to situations. Situations can be seen as consequences of actions and for this reason people can define rules about acceptability of typical actions. Finally, socio-ethical discourse create general ethical rules. Intelligent systems can provide important support in ethical process and thus the difference between weak and strong ethical AI is polar.

show abstract