Discussion on Machine Learning

AI Discrimination of Machine Learning

The artificial intelligence is increasingly used in recruiting and might discriminate accidentally against minorities as well as women if the data is infective and ambiguous. furthermore. It may sue the AI vendors for this kind of discrimination along with employers. The artificial intelligence is able to effectively provide, rational as well as consistent assessments. However, the decision making of the machines through algorithm has proven discrimination potentially (Aitkenhead, Dalgetty, Mullins, McDonald, & Strachan, 2003).

Unintended discrimination behaviors in humans: Different types of discrimination behaviors in humans are identified. After researching human behaviors, it was determined that the unintended discrimination in humans vary but some notable discriminations can be: age, disability, reassignment of gender, and race.

Discrimination of AI Model: There is some key discrimination in the artificial intelligence (AI) model. The artificial intelligence is the source of bother skepticism and enthusiasm in different ways. AI is no more limited to the innovation labs but some discrimination can be seen in the AI Models. The first discrimination is data security and privacy. Most of the AI-based machines depend on high volume data for learning and decision making intelligently. These models can be sensitive as well as personal in nature for learning and enhancing themselves which makes it vulnerable to serious issues such as identity theft and data breach.

Another inherent problem with the AI models is that they can only be good or as bad based on the data. The bad data is often associated with racial, communal, gender, or ethnic biases. There is a risk or threat of showing inappropriate results when the AI system unable or misses the mark on racial sensitivity (Rao, Monkowski, & Roukos, 1995).

Furthermore, at the time of making a deep learning model is to make a decision what they actually want to obtain. The problem is that the decisions are made for many business reasons rather than any kind of fairness and discrimination.

Victims of discrimination: The rapid development of AI has seen in the past decades and resulted in the reliance and extensive usage in several fields that influence daily lives and human rights. By researching further on the discrimination, it is determined that some companies are facing some challenges such as data labeling, case-specific learning, biasing, lack of understanding of machine learning among non-technical employees, and some legal issues (Yavuz, 2019).

Preventions: In the preventions, some important preventions are described in this document which is given below.

·         Using the representative dataset

·         Choosing the right AI model and algorithm

·         Effectively monitor and review of outputs generated from the machine.

References of Machine Learning 

Aitkenhead, M. J., Dalgetty, I. A., Mullins, C. E., McDonald, A. J., & Strachan, N. J. (2003). Weed and crop discrimination using image analysis and artificial intelligence methods. Computers and electronics in Agriculture, 157-171.

Rao, P. S., Monkowski, M. D., & Roukos, S. (1995). Language model adaptation via minimum discrimination information. In 1995 International Conference on Acoustics, Speech, and Signal Processing, 161-164.

Yavuz, C. (2019). Machine Bias Artificial Intelligence and Discrimination. Master of Laws in International Human Rights Law.