Machine learning’s limitations in avoiding automation of bias

Varona, Daniel; Lizama-Mué, Yadira; Suárez, Juan Luis

doi:10.1007/s00146-020-00996-y

Cited by 14 publications

(9 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We observed biased predictions due to imbalance news items in few cases (e.g. the model predicted fake news items with higher accuracy than real news items) and therefore this example presents a scenario of limitation of ML algorithms in avoiding automation of bias [31].…”

Section: Resultsmentioning

confidence: 86%

Predictive linguistic cues for fake news: a societal artificial intelligence problem

Aneja

Kumaraguru

2022

IJ-AI

View full text Add to dashboard Cite

<span>Media news are making a large part of public opinion and, therefore, must not be fake. News on web sites, blogs, and social media must be analyzed before being published. In this paper, we present linguistic characteristics of media news items to differentiate between fake news and real news using machine learning algorithms. Neural fake news generation, headlines created by machines, semantic incongruities in text and image captions generated by machine are other types of fake news problems. These problems use neural networks which mainly control distributional features rather than evidence. We propose applying correlation between features set and class, and correlation among the features to compute correlation attribute evaluation metric and covariance metric to compute variance of attributes over the news items. Features unique, negative, positive, and cardinal numbers with high values on the metrics are observed to provide a high area under the curve (AUC) and F1-score.</span>

show abstract

Section: Resultsmentioning

confidence: 86%

Predictive linguistic cues for fake news: a societal artificial intelligence problem

Aneja

Kumaraguru

2022

IJ-AI

View full text Add to dashboard Cite

show abstract

“…Both the engineering approach and the stipulation of the regulatory approach need to be incorporated into an integrative mechanism oriented to reducing and mitigating algorithmic decision-making (ADM) systems-produced discriminatory outcomes, analyzed in previous studies [1,2]. The traditional approach conducted to manage discrimination, prejudice or bias, and algorithmic unfairness historically exhibits a reactive character that must be overcome, as criticized in [3]. Additionally, the needed proactive approach must incorporate the determination of possible remedy actions due to discriminatory ADM systems' outcomes.…”

Section: Introductionmentioning

confidence: 99%

Discrimination, Bias, Fairness, and Trustworthy AI

Varona¹,

Suárez²

2022

Applied Sciences

Self Cite

View full text Add to dashboard Cite

In this study, we analyze “Discrimination”, ”Bias”, “Fairness”, and “Trustworthiness” as working variables in the context of the social impact of AI. It has been identified that there exists a set of specialized variables, such as security, privacy, responsibility, etc., that are used to operationalize the principles in the Principled AI International Framework. These variables are defined in such a way that they contribute to others of more general scope, for example, the ones studied in this study, in what appears to be a generalization–specialization relationship. Our aim in this study is to comprehend how we can use available notions of bias, discrimination, fairness, and other related variables that will be assured during the software project’s lifecycle (security, privacy, responsibility, etc.) when developing trustworthy algorithmic decision-making systems (ADMS). Bias, discrimination, and fairness are mainly approached with an operational interest by the Principled AI International Framework, so we included sources from outside the framework to complement (from a conceptual standpoint) their study and their relationship with each other.

show abstract

“…Ethically, one of the central issues in AI ethics is: how can they be created in such a way that they do not share the same ethical biases as their creator [8,41]. When Google's word embedding tool was used to solve verbal analogy problems, researchers found that it performed the task with blatant and rampant gender bias [41,65]. AI bias has the potential to be a particular problem in fields prone to unequal treatment of individuals such as medicine or fields that have the potential to have severe environmental impacts if the AI involved are only concerned with maximum yields, for example agriculture [30,32,66].…”

Section: Introductionmentioning

confidence: 99%

Ethical, legal, social, and economic (ELSE) implications of artificial intelligence at a global level: a scientometrics approach

et al. 2022

View full text Add to dashboard Cite

Artificial intelligence (AI) development and applications are growing rapidly. Simultaneously, researchers have also been exploring the ethical, legal, social, and economic (ELSE) implications of it. However, global mapping of the ELSE implications of AI is lacking, hence we explored it through mixed qualitative and quantitative research methods. Using a scientometrics analysis of the publication records (between 1991 and 2020; n = 1028), and content analysis of highly cited publications, our study provided insights on the ELSE implications of AI. Our study findings indicate that ELSE implications of AI development started gaining momentum globally over the last 5 years and we predict that by the end of this decade publication numbers will be more than 750 per year. Europe (46%) and North America (33%) were leaders in publications in this area while Africa (1.8%) and South America (1.4%) have lagged behind. Additionally, the computer science (350) research area had the maximum number of ELSE implications of AI publications, followed by humanities and social sciences (e.g., legal, policy; 322), but have not been explored extensively in the agricultural sciences (23). We observed that the major disparities in studies of ELSE implications of AI were found to be a combination of economics, governance, sociocultural, and policy factors. ELSE implications must be explored through a multidisciplinary approach, taking into consideration the stakeholders' perspectives right at the inception of AI systems development to gain trust and better adoption by the end users. KeywordsAI and ethics • Scientometrics • Keyword analysis • ELSE implications of AI • Co-occurrence analysis • AI ethics and global mapping • Content analysis • AI ethics and agricultural sciences * Debasmita Patra

show abstract

Machine learning’s limitations in avoiding automation of bias

Cited by 14 publications

References 11 publications

Predictive linguistic cues for fake news: a societal artificial intelligence problem

Predictive linguistic cues for fake news: a societal artificial intelligence problem

Discrimination, Bias, Fairness, and Trustworthy AI

Ethical, legal, social, and economic (ELSE) implications of artificial intelligence at a global level: a scientometrics approach

Contact Info

Product

Resources

About