04/23/23

AI’s Impact on Human Identity & vice-versa

notes on Inês Hipólito’s paper deploying complex systems theory for responsible AI

If you’ve been following along, you know that I am devoted to studying complex adaptive systems. You can read my previous pieces on  why I study complexity, my thoughts on biotech x complexity, and the limit to the machine metaphor. 

Today I came across this paper by Inês Hipólito shared via Barton on Bluesky, I felt like someone had gone inside my brain and articulated exactly how I have been thinking about responsible AI and inductive bias.

My hope is for everyone to read and understand this. Below are some notes and my thoughts on the paper. 

The following is my exploration of Hipólito’s paper’s commentary of Schneider’s book.

The book is Susan Schneider’s Artificial You
The paper is The Human Roots of Artificial Intelligence by Inês Hipólito 

TLDR of book

Schneider’s book explores the possibility of creating conscious AI and the ethical and philosophical implications that arise from it. Schneider's argument highlights the potential for AI to significantly transform our self-perception and challenge conventional ideas of identity and consciousness. This, in turn, prompts significant ethical and cognitive science considerations that must be approached with caution.

TLDR of paper

Hipólito’s paper argues that AI cannot be viewed as an independent entity from human social and cultural contexts, presenting a critique of realism with regard to AI. 

In other words, the argument suggests that AI does not exist independently of human practices and values, but rather emerges from them.

Simone de Beauvoir's "Ethics of Ambiguity" explores the relationship between freedom, responsibility, and ethical action in a world that is uncertain and ambiguous. She argues that human beings must embrace this ambiguity and create their own meaning in life instead of relying on external authorities or fixed moral codes. She suggests that ethical action involves recognizing the ambiguity of our situation and making choices that affirm our own freedom and the freedom of others. Her work has influenced feminist and existentialist philosophy as well as political and social thought.

The paper proposes that AI design should be considered a cultural practice, and instead of advocating for realism about AI's independent agency, it suggests using De Beauvoir's Ethics of Ambiguity to provide specific guidelines for designing AI in an ethical manner. This approach recognizes the influence of culture on AI development and emphasizes the importance of ethical considerations that prioritize self-determination.

Paper’s Arguments

#1 Rejecting Realism about AI 

  • Assumption that AI is inherently neutral and objective (the realism perspective)  is dangerous 

  • The reality is that AI systems are shaped by the cultural and social contexts in which they are created and deployed, and overlooking this risks reinforcing existing power structures and inequalities

  • Rather than viewing AI as an autonomous entity, we should recognize it as an emergent practice rooted in human sociocultural dynamics

  • Doing so acknowledges the imprint of sociocultural forces on AI and emphasizes the importance of ethical considerations in AI design

#2 Human Cultural Roots of AI 

  • The philosophical importance of understanding how AI is rooted in human sociocultural practices, which are shared behaviors, beliefs, and customs that characterize a given society or culture. 

  • The development of AI reflects the sociocultural forces that have shaped and molded it, as well as the human agents that have been the driving force behind its creation. 

  • The design of AI embodies the values and biases of a given sociocultural context and is contingent upon cultural practices that dictate which issues are deemed worthy of attention and how values are operationalized in the design process. 

  • Proper evaluation of AI's meaning and value necessitates a thorough consideration of the social and cultural dynamics that nurtured it. 

  • The core problem of how to assign predetermined fixed rules or principles to regulate AI in the complexity of the world and unique circumstances of each situation is a delicate balance between autonomy and freedom.

#3 Agency Amongst others 

  • The interconnectedness of human beings shapes our self-understanding, and agency is closely tied to our ability to make choices for ourselves. 

  • AI can limit our agency by restricting our choices, perpetuating power imbalances and inequalities, and replacing human decision-making with automated processes. 

  • The implementation of AI holds the potential to either reinforce or challenge social segregation and promote well-being, but raises ethical concerns regarding accountability, transparency, and the potential loss of human agency. 

  • The tension between freedom and responsibility can inform AI design.

#4 Ethics of Ambiguity for Ethical AI Design 

  • There is inherent ambiguity in ethical decision-making and it’s important to recognize the autonomy and agency of others. 

  • AI cannot rely on predetermined fixed rules or principles to guide actions due to the complexity of the world and unique circumstances of each situation. 

  • Ethical decision-making is complicated due to navigating conflicting values, beliefs, and interests. 

  • This part outlines three main reasons why ethical decision-making is difficult: the plurality of values and ethical theories, the situatedness of ethical decision-making, and the uncertainty and unpredictability of the future.

Active agents shaping their lives with others 
The Ethics of Ambiguity (De Beauvoir) highlights the importance of human agency and autonomy in shaping their own lives, which can inform AI design by prioritizing user control and consent, addressing potential biases in data sources, and continuously reflecting on underlying assumptions and values in the design process. 

Designers should create AI systems that empower users to make their own choices and determine their own values, while also being transparent about data collection and usage. 

State of ambiguity with others 
The design and implementation of AI should take into account the evolving nature of human existence through choices and interactions with others. The potential impact of AI on individuals and society must be deeply understood and reflected in the design process, with consideration of relationships and user autonomy. AI systems should promote positive social outcomes and incorporate values of empathy, solidarity, and social justice to mitigate negative effects and avoid creating AI in a vacuum.

Self-reflection and critical awareness
The role of ethics in AI design is crucial in ensuring that the technology aligns with societal values and contributes to a just and equitable world. To achieve this, designers must engage in critical reflection, acknowledge their biases and assumptions, and consider the potential positive and negative consequences of their technology. Ongoing ethical reflection and dialogue are necessary to ensure that the technology aligns with societal values and contributes to a more just and equitable world. Ethical decision-making is inherently ambiguous and requires consideration of complex and often conflicting values, beliefs, interests, and contexts.

My thoughts

Complex systems are made up of numerous interconnected interacting parts. The humans building these algorithms have biases. The future is unpredictable. Ethical decision-making is contextual. The development of an AI system reflects the values and prejudices of a specific sociocultural environment and is dependent on cultural norms that determine which problems are considered important and how principles are put into practice during its design phase. 

How do we consider complex and conflicting values systems, morals, beliefs, contexts while designing systems? 

If we create AI systems that let users make their own choices and determine their own values – we make this aspect an inherent part of the deployment of an algorithm, rather than a one-size-fits-all, which doesn’t really fit “all” because “all” are not the same. 

This is all rooted in the basic fact that the behaviour of complex systems is unpredictable and emergent. The AI system’s design must be adaptable and sustainable to emergence from interactions between the different aspects of said system, and co-evolve with it. 


Philosophy Glossary

Realism: Realism is a broad term that refers to the artistic or literary representation of things as they actually are, without idealization or exaggeration. It emerged in the 19th century as a reaction against Romanticism, which celebrated emotion and imagination over reason and logic. Realism sought to capture the realities of everyday life, often focusing on the struggles of ordinary people in society.

Self Determination: Self-determination is the ability of an individual or group to make decisions and choices for themselves without external influence or coercion. It is the power to govern one's own life and destiny, and to have control over one's own actions and decisions. In political contexts, self-determination often refers to the right of a people to determine their own political status and form of government. In a more general sense, self-determination is the idea that individuals should have the freedom to live their lives according to their own values and beliefs, as long as they do not infringe on the rights of others.

This exploration is 23/50 of my 50 days of learning. Subscribe to hear about new posts.