Algorithms and Bias in Decision-Making by Caleb Nielsen

How algorithms can have biases

caleb22n — 2025-03-23 21:42:36 UTC

Algorithms are used in many different industries, most notably in artificial intelligence. Bias happens in algorithms when data is collected. The data can be incomplete, misused, mislabeled, or even misrepresentative

Significance

caleb22n — 2025-03-23 21:51:12 UTC

Algorithms and biases are things that go hand in hand in today's world. This is because many people directly use or interact with algorithms in some way every single day. Social media, online shopping, job searching, anything you do online has some sort of algorithm built in.

Summary

caleb22n — 2025-03-24 00:40:01 UTC

Ehtically algorithms have many upsides if done correctly, but equally as many downsides if done incorrectly.

Pros include increased efficiency, fast training, saves energy.

Cons include the many ways biases can be introduced to algorithms. skewed data, feedback loop, incomplete training data, and biased training data.

What can biases lead to?

caleb22n — 2025-03-24 01:02:37 UTC

Data misrepresentation is one example of what can happen when an algorithm bias occurs. Data misrepresentation happens when the coding on the algorithm is pointed intentionally or unintentionally regardless of what the data statistically portrays. Many times this is used by people to sway people towards their agenda.

Real world example Amazon

caleb22n — 2025-03-24 01:08:29 UTC

It was recently discovered that Amazon's hiring algorithm favored men over women. This created an unequal opportunity for candidates and was clear discrimination against female applicants. It was discovered that they trained the algorithm with past employee resumes who were primarily male.

Real world example facial recognition

caleb22n — 2025-03-24 01:16:54 UTC

Facial recognition has been found to have higher error rates in recognizing people of color. While white people have been found to have a much higher percentage of success.

Real world example credit scoring models

caleb22n — 2025-03-24 01:22:21 UTC

Research has shown that people of color are not represented fairly in starting their credit journey. This is because it uses previous historical data that is already biased.

What can be done to regulate?

caleb22n — 2025-03-24 01:31:30 UTC

Regulation and review are important for doing anything in life. It allows products and services to become better through critique and evaluation. To regulate biases in algorithms there needs to be longer testing periods with the algorithms and data needs to be presented as is with no blemishes. Lawmakers should also step in and make companies clarify how often they can be used and make users aware of when they are being used.

Real world example Illinois AI video Act

caleb22n — 2025-03-24 01:35:34 UTC

In 2019 Illinois required employers to notify applicants that AI was being used to judge their video interviews

Real world example New York City

caleb22n — 2025-03-24 01:39:52 UTC

In 2021 New York City passed local law 144 which required companies to use AI in the hiring process to conduct audits on the algorithms they are run on.

What are the consequences of companies relying on algorithms?

caleb22n — 2025-03-24 01:49:48 UTC

There are both good and bad consequences of companies relying on algorithms.

The good include cost reduction, increased efficiency, and data-driven insights and results.

The bad include unfair hiring practices, financial discrimination, and a lack of transparency and accountability.

Real world Example Washington DC school district

caleb22n — 2025-03-24 01:53:37 UTC

They created an algorithm to fair the lowest-performing teachers in the district. The algorithm sought to fire around 200 teachers in that district many of whom received many honors for their high level of teaching

Real world example Meta misinformation

caleb22n — 2025-03-24 02:00:29 UTC

From the years 2016-2020, Meta was found guilty of fixing their algorithms on Facebook to promote misinformation in an attempt to influence the election and sway public opinion.

Business benefits and risk

caleb22n — 2025-03-24 02:19:06 UTC

Algorithms offer businesses many ways to streamline and utilize data like never seen before in human history. However, companies need to be aware of they way they use them and how much they use them. Humans are complex and intricate creatures that can't be boiled down to just a number.

caleb22n — 2025-03-24 02:19:18 UTC

consumers benefits and risks

caleb22n — 2025-03-24 02:25:39 UTC

Consumers just like businesses can enjoy many benefits from algorithms as well like the personalization of data and the ability to make decisions based on data faster and more efficiently. However, algorithms unchecked can cause discrimination, lack of transparency, overreliance, and a decrease in critical thinking skills.

caleb22n — 2025-03-24 02:26:04 UTC

sources

caleb22n — 2025-03-24 02:28:27 UTC

ESLBuzz. (n.d.). Plural of bias. ESLBuzz. https://eslbuzz.com/plural-of-bias/
The Ladders. (2018, October 10). Amazon reportedly scraps AI recruiting tool biased against women. The Ladders. https://www.theladders.com/career-advice/amazon-reportedly-scraps-ai-recruiting-tool-biased-against-women
McGill, M. (2018, July 26). Amazon's facial recognition shows bias against people of color. Axios. https://www.axios.com/2018/07/26/amazon-facial-recognition-racial-bias
DeWees, B. (2023). Why businesses need 3rd party algorithm audits. BDeWees.com. https://www.bdewees.com/business-needs-3rd-party-audit/
AP News. (2021, October 26). Facebook knew it was polarizing, but let it happen anyway. AP News. https://apnews.com/article/facebook-instagram-polarization-misinformation-social-media-f0628066301356d70ad2eda2551ed260
Stocksy. (n.d.). Teacher helping her young pupils in a class activity [Photograph]. Stocksy. https://www.stocksy.com/photo/2433618/teacher-helping-her-young-pupils-in-a-class-activity
TEDx Talks. (2019, March 18). How AI can enhance our memory, work, and social lives [Video]. YouTube. https://www.youtube.com/watch?v=UG_X_7g63rY
CNBC Television. (2022, April 14). The impact of AI bias on hiring and business [Video]. YouTube. https://www.youtube.com/watch?v=og67qeTZPYs
Vox. (2020, August 17). How AI-powered facial recognition is changing our world [Video]. YouTube. https://www.youtube.com/watch?v=y6FDBDGNYFc