1 of 28

Ethical Dilemmas

in AI

Navigating Implementation in our Schools

July 22, 2025

bit.ly/aiDilemma2025

2 of 28

Welcome

Marissa McNamara

Educational Technology Specialist

mmcnamar@btboces.org

What brought you to this session?

3 of 28

Define algorithmic bias and its implications

Explore and apply a critical lens to ethically complex scenarios

Identify applications and next steps for decision-making shifts

Outcomes

bit.ly/aiDilemma2025

4 of 28

What is bias?

In thinking about the human brain, bias is an “adaptive process that allows us to use prior knowledge and experiences to inform our decisions and actions in the present.”

5 of 28

Bias in AI

Programmer Bias

Training Data Bias

6 of 28

Bias in AI

Programmer Bias

Programmers fall within the dominant culture.

7 of 28

Bias in AI

Training Data Bias

“A set of labeled examples used to teach AI systems and machine learning models to recognize patterns, make decisions, and improve over time.”

  • From Google’s Gemini

Example: Using Google Image Search type in "professional hairstyles" then search for "unprofessional hairstyles."

8 of 28

Bias in AI

Algorithmic Bias

Programmer Bias

Training Data Bias

9 of 28

Algorithmic Bias-

a Definition

The systemic and unfair discrimination in AI algorithms due to biases in the training data or the design of the algorithms themselves, leading to unequal treatment of different groups based on characteristics like race, gender, or

socioeconomic status.

10 of 28

Who is the most impacted by algorithmic bias?

Everyone is impacted, but we especially need to consider folks who are outside the dominant culture in the United States and around the World:

  • Black and Indigenous People of Color (BIPOC)
  • Members of the LGBTQIA+ Community
  • Women and Gender-Nonconforming Individuals
  • Children and the Elderly
  • Members of non-Christian religions
  • Non-Americans
  • Disabled Folks
  • People living in Poverty
  • Non-English Speaking Folks

*This list is not exhaustive

11 of 28

Keeping Humans

in the Loop

This graphic was created with Napkin.ai

12 of 28

How might algorithmic bias uphold existing systems of inequity and bias in our society, especially in schools?

Turn and Talk

13 of 28

We are at a turning point

14 of 28

Ethical Dilemmas

How can we optimize AI's beneficial impact while reducing risks and adverse outcomes?

15 of 28

Tug of War

Two or more conflicting options- You can't choose both (or all).

Significant consequences- The outcome matters.

No clear "right" answer- Every option has downsides.

Moral or practical tension- Challenges to your values, priorities, or goals.

16 of 28

Ethical Dilemmas

In small groups:

  • Work through the dilemma scenarios on slides 17-26 using
  • Identify a note taker & time keeper
    • Note Taker will jot thoughts on Padlet
    • Time Keeper will help keep pace�about 6 minutes per scenario

bit.ly/aiDilemma2025

17 of 28

#1

18 of 28

#2

19 of 28

#3

20 of 28

#4

21 of 28

#5

22 of 28

#6

23 of 28

#7

24 of 28

#8

25 of 28

#9

26 of 28

#10

27 of 28

Ethical Decision Making

Frameworks & Resources

28 of 28

"Am I erecting or dismantling barriers?"

What is one idea that you want to hold on to as we begin a new school year?

In closing…