Challenging dominance-based AI metaphors in design is crucial to ensure that AI systems are built with a more inclusive, empathetic, and human-centered approach. Dominance-based metaphors, often drawn from hierarchical power structures, can reinforce oppressive social dynamics and alienate users. Here are several ways to challenge and reframe these metaphors in design:
1. Emphasize Collaboration Over Control
Dominance-based metaphors often frame AI as a tool of control or surveillance, such as “master” and “slave” or “controller” and “user.” These metaphors suggest an imbalance of power. To challenge this:
-
Use Co-agency Models: Frame the relationship between AI and the user as a partnership, where both contribute to a decision-making process. Metaphors like “companion,” “advisor,” or “co-pilot” suggest that the AI is supporting the user rather than controlling them.
-
Shared Control: Focus on shared agency, where users have the ultimate decision-making power, and the AI acts as an enabler. This can be reflected in interface designs that emphasize user control and autonomy.
2. Promote Ethical and Empathetic AI Systems
Dominance-based metaphors can reinforce a lack of empathy by treating AI as a tool of efficiency or even exploitation.
-
Design for Empathy: Use metaphors that highlight the AI’s role in understanding, responding to, and supporting human emotions. For example, “listener,” “supporter,” or “guide” convey a sense of care and attention.
-
Healing Metaphors: For emotionally charged or sensitive interactions, using metaphors like “healer,” “comforter,” or “empath” can help AI systems to be perceived as more compassionate.
3. Reject Militarized Language
Many AI designs draw on militarized metaphors, such as “command,” “attack,” “defense,” or “battle.” These metaphors contribute to the perception of AI as something inherently aggressive or combative.
-
Use Peaceful Language: Instead, frame AI in terms of cooperation and understanding. Words like “assist,” “help,” “navigate,” or “facilitate” avoid aggressive connotations and can reshape the narrative around AI as a tool for peaceful collaboration.
-
Avoid Conflict-Driven Logic: Where appropriate, remove language and logic that pits AI against human decision-making, and instead focus on how AI can enhance human capabilities in a harmonious way.
4. Encourage Intersectional and Inclusive Design
Dominance-based metaphors can often be exclusionary, centering the design around specific user groups and power structures. To ensure more inclusive design, it’s important to:
-
Reflect Diverse Perspectives: Involve diverse voices, especially those from marginalized communities, in the AI development process. This ensures that the metaphors used resonate across different cultures, identities, and experiences.
-
Contextual Sensitivity: Recognize that metaphors should be adapted based on the context in which AI is used. A metaphor that works in one context (e.g., healthcare) might not be appropriate in another (e.g., social justice). Using flexible and context-aware metaphors helps to challenge dominant narratives.
5. Humanize AI Systems
Rather than positioning AI as an all-powerful or omniscient force, challenge the hierarchical metaphors by humanizing AI systems.
-
Personification with Caution: While personifying AI can sometimes be problematic if it leads to unrealistic expectations, using non-dominant, human-centered metaphors can help shift the perception of AI from an abstract, detached entity to a system with specific, clear, and supportive roles.
-
Non-Dominant Imagery: Using metaphors based on nature, care, and mutual growth can counterbalance the traditional “tool” metaphor. For example, a metaphor like “gardener” (for AI systems that help cultivate knowledge or creativity) implies a nurturing, supportive role.
6. Foster Transparency and User Agency
Metaphors that imply dominance often obscure the inner workings of AI systems, leading to mistrust and fear.
-
Transparency in Communication: Use metaphors that communicate openness and clarity. For instance, “window” (to the data or logic of AI) or “map” (showing the possible paths AI can take) emphasizes transparency and allows users to better understand the AI’s decision-making processes.
-
Decision Support: Frame AI as a system that enhances human judgment rather than making decisions independently. Words like “assistant,” “guide,” or “navigator” suggest that the AI is there to augment human agency, not replace it.
7. Use Metaphors that Encourage Ethical AI Development
A key part of challenging dominance-based AI metaphors is embedding ethical principles into the design process.
-
Responsibility and Accountability: Metaphors that focus on accountability, like “steward,” “guardian,” or “custodian,” emphasize that the creators and operators of AI systems are responsible for ensuring they operate ethically and in the best interests of users.
-
Sustainability: Metaphors like “caretaker” or “nurturer” suggest that the AI system is not only about meeting immediate needs but is designed with a long-term, sustainable approach, taking into consideration the environment, society, and user well-being.
8. Leverage Speculative Design to Reimagine Dominance
Speculative design can challenge existing metaphors and provide alternative futures for AI interactions.
-
Reframe Power Dynamics: Speculative design can help imagine new power dynamics and metaphors, such as “circles” or “networks” instead of hierarchies. This can promote a more equitable distribution of agency and control in AI interactions.
-
Engage Diverse Imaginaries: By using speculative scenarios, designers can push boundaries and envision AI as something other than a tool of dominance. Exploring what an AI system designed around community, care, or shared goals might look like could shift the metaphorical landscape.
Conclusion
To challenge dominance-based AI metaphors in design, it’s necessary to rethink the relationship between AI and the user from one of power and control to one of collaboration, empathy, and mutual respect. By adopting more inclusive, transparent, and ethical metaphors, AI systems can become more human-centered and supportive of diverse users, fostering trust and encouraging positive social impacts.