Rethinking 'the Human' in AICEFR B1
16 Apr 2026
Adapted from Guest Contributor, Global Voices • CC BY 3.0
Photo by Steve A Johnson, Unsplash
Kira traces the problem to a long history in which animal, ecological, ancestral and spiritual forms of knowledge were dismissed. She argues that the narrow idea of “the human” has often been shaped by white supremacy, patriarchy, ableism and cisheteronormativity, and that reaffirming that idea in AI can reproduce exclusion rather than undo it.
She also discusses how the border between natural and artificial creates moral rules about who counts as embodied or real, and how those rules have been used to other people with marked bodies, genders or abilities. Kira describes a growing crisis of the real: AI images and voices now compete with reality, creating many images but making verification harder.
Her practical proposals include building situated datasets grounded in relationships, consent and accountability, and imagining smaller, slower models tuned to communities. She insists imagination should go with dismantling existing harms.
Difficult words
- dismiss — refuse to accept as important or truedismissed
- white supremacy — belief that white people are superior and dominate
- patriarchy — social system where men hold primary power
- ableism — discrimination against people with disabilities
- cisheteronormativity — belief that cisgender heterosexual identities are normal
- embodied — having a physical body or physical form
- verification — process of checking whether something is true
- dataset — a collection of organized data for analysisdatasets
- consent — permission given for something to happen
Tip: hover, focus or tap highlighted words in the article to see quick definitions while you read or listen.
Discussion questions
- Have you seen examples where online images or voices were hard to trust? Describe one briefly.
- What small steps could researchers take to include community voices when they build AI?
- How important is consent when people’s images or voices are used? Give one reason.
Related articles
LLMs change judgments when told who wrote a text
Researchers at the University of Zurich found that large language models change their evaluations of identical texts when given an author identity. The study tested four models and warns about hidden biases and the need for governance.
MLB Toolbox: Money and Baseball Performance
Boston University finance expert Mark Williams and colleagues created MLB Toolbox, a data platform that compares team payrolls with on-field performance. The site ranks teams by spending efficiency, wins and WAR and lets users test roster choices.
AI and Wearable Devices for Type 2 Diabetes
A meta-review from the University at Buffalo examines AI-enhanced wearable devices for Type 2 diabetes and prediabetes. The study finds predictive benefits and important limits, and calls for larger, more transparent studies before routine clinical use.
