A carregar...
Publicação
Artificial intelligence and discrimination. How artificial intelligence impacts on the right to non-discrimination in the law enforcement environment
| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 711.8 KB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
A Inteligência Artificial (IA) pode reforçar os preconceitos, amplificando ou
perpetuando o preconceito enraizado na sociedade. No caso do policiamento preditivo,
os algoritmos baseados em IA podem levar a previsões tendenciosas baseadas em (por
exemplo) ações policiais anteriores, o que pode resultar no policiamento excessivo de
alguns grupos e numa aplicação desigual da lei. Para assegurar a utilização justa dos
sistemas de IA, encorajar a responsabilização e a transparência, e prevenir danos às
pessoas desfavorecidas, é essencial compreender os possíveis efeitos da IA na
discriminação. Este é o primeiro passo para desenvolver práticas éticas e inclusivas de
IA, identificando e mitigando potenciais preconceitos nos sistemas de IA, aumentando
a confiança pública na tecnologia de IA e, por último, o avanço de aplicações
responsáveis e justas de IA em vários domínios, incluindo o policiamento. De forma
geral, esta tese de mestrado contribuirá para a discussão em curso sobre o
desenvolvimento e utilização responsável da IA.
Artificial Intelligence (AI) can reinforce prejudice by amplifying or perpetuating societal bias. In the case of predictive policing, AI-based algorithms may lead to biased predictions based on (for instance) prior police action, which might result in the overpolicing of some groups and uneven application of the law. To ensure the fair and just use of AI systems, encourage accountability and transparency, and prevent harm to disadvantaged people, it is essential to comprehend the possible effects of AI on discrimination. This is the first step towards developing ethical and inclusive AI practices, identifying, and mitigating potential biases in AI systems, enhancing public trust in AI technology and, lastly, the advancement of responsible and fair AI applications in various domains, including policing. Overall, this master thesis will contribute to the ongoing discourse on the responsible development and use of AI.
Artificial Intelligence (AI) can reinforce prejudice by amplifying or perpetuating societal bias. In the case of predictive policing, AI-based algorithms may lead to biased predictions based on (for instance) prior police action, which might result in the overpolicing of some groups and uneven application of the law. To ensure the fair and just use of AI systems, encourage accountability and transparency, and prevent harm to disadvantaged people, it is essential to comprehend the possible effects of AI on discrimination. This is the first step towards developing ethical and inclusive AI practices, identifying, and mitigating potential biases in AI systems, enhancing public trust in AI technology and, lastly, the advancement of responsible and fair AI applications in various domains, including policing. Overall, this master thesis will contribute to the ongoing discourse on the responsible development and use of AI.
Descrição
Palavras-chave
Artificial intelligence Human rights Discrimination Right to nondiscrimination Predictive policing Regulation