AI ETHICS & BEHAVIORAL SCIENCES
What Does 'Human-Centred AI' Mean?
While it seems sensible that human-centred artificial intelligence (AI) means centring "hu-man behaviour and experience,” it cannot be any other way. AI, I argue, is usefully seen as a relationship between technology and humans where it appears that artefacts can perform, to a greater or lesser extent, human cognitive labour. This is evinced using examples that juxtapose technology with cognition, inter alia: abacus versus mental arithmetic; alarm clock versus knocker-upper; camera versus vision; and sweatshop versus tailor. Using novel definitions and analyses, sociotechnical relationships can be seen as varying types of: displacement (harmful), enhancement (beneficial), and/or replacement (neutral) of human cognitive labour. Ultimately, all AI implicates human cognition; no matter what. Obfus-cation of cognition in the AI context—from clocks to artificial neural networks—results in distortion, in slowing critical engagement, perverting cognitive science, and indeed in limiting our ability to truly centre humans and humanity in the engineering of AI systems. To even begin to de-fetishise AI, we must look the human-in-the-loop in the eyes.
Executive Impact Overview
This analysis redefines 'Human-Centred AI' by dissecting its relationship with human cognitive labour. It reveals that AI can displace, enhance, or replace human tasks, with significant implications for skill development, labour transparency, and societal well-being. Focusing on the 'human-in-the-loop' is crucial to avoid harmful obfuscation and ensure ethical, human-centric development of AI systems. The article challenges correlationist views and calls for a critical engagement with AI's true impact on human cognition and society.
0
Years of AI Hype
0
Types of AI Relationship
0
HCAI Themes Challenged
Deep Analysis & Enterprise Applications
Select a topic to dive deeper, then explore the specific findings from the research, rebuilt as interactive, enterprise-focused modules.
Proposed Two-Step Redefinition of AI
The radical redefinition of AI involves two main steps: first, discerning if a relationship between an artefact and human cognition exists; second, characterizing this relationship as replacement, enhancement, or displacement of cognitive labour.
Discern Relationship: Artefact to Cognition
→
Characterise Relationship: Labour Type
→
Determine Type: Replacement, Enhancement, or Displacement
| Property | Replacement | Enhancement | Displacement |
|---|---|---|---|
| Valence | Neutral | Beneficial | Harmful |
| Effect on Cognition | Unaffected | Reskilling | Deskilling |
| Labour Obfuscation | Minimal | Possible | Maximal |
| Human Equivalence | Worse or Same | Different | No |
| Human-in-the-Loop | Rare | Possible | Common |
| Human Input | Transparent | Transparent | Opaque |
| Desired Output | Specified | Formal | Unspecified |
Key Factor: Labour Obfuscation
Maximal Labour Obfuscation in Displacement AIA critical dimension in understanding AI's impact is the extent to which human cognitive labour is hidden or made transparent. Displacement AI relationships are often characterized by maximal obfuscation, masking the true human effort.
Abacus: An Enhancement for Mental Arithmetic
The abacus aids mental arithmetic, not replacing the underlying skill but enhancing it, promoting new cognitive skills. It requires full human cognition for operation, with minimal labour obfuscation.
- Enhancement AI: The abacus provides marked benefits for mental arithmetic skills.
- Reskilling: Users develop new skills, fostering a symbiotic relationship.
- Transparent Input: Requires 'full-blown cognition' from the user, ensuring active engagement.
- Low Obfuscation: The human effort is evident, promoting understanding rather than hiding it.
| Property | Calculator vs. Human |
|---|---|
| Label | Replacement |
| Valence | Neutral |
| Effect on Cognition | Unaffected (for skilled adults) |
| Labour Obfuscation | Moderate |
| Human Equivalence | Better (speed/accuracy) |
| Human-in-the-Loop | None |
| Human Input | Function, numbers |
Digital Computer: Displacement of Human Labour
Historically, human computers (often women) performed complex calculations. The rise of digital computers led to their displacement, devaluing their cognitive labour and creating a harmful relationship, often characterized by obfuscation and lack of recognition.
- Displacement AI: Digital computers took over tasks from highly skilled human computers.
- Deskilling: Original human computation skills were marginalized.
- High Obfuscation: The electronic computer often hid the intellectual labour involved in programming and maintenance.
- Harmful Valence: Led to the devaluation and erasure of human labour.
| Property | Alarm Clock vs. Knocker-Upper |
|---|---|
| Label | Replacement |
| Valence | Neutral |
| Effect on Cognition | Reskilling (managing time) |
| Labour Obfuscation | Minimal |
| Human Equivalence | Worse or Same (no guarantee) |
| Human-in-the-Loop | None (after initial setting) |
| Human Input | Current time, ring time, energy |
Camera: Enhancement for Human Vision and Memory
The camera enhances human vision and memory without deskilling. It allows us to capture and recollect moments, augmenting our ability to perceive and recall, though user input (framing, focus) is essential.
- Enhancement AI: Cameras augment human visual capabilities and memory.
- Unaffected Cognition: Does not negatively impact the ability to visually perceive.
- User-in-the-Loop: Requires active human framing, memory, or film/energy source for desired output.
- Beneficial Valence: Helps recollect memories, enhancing human ability to think about the past.
Sweatshop Labour Obfuscation
Maximal Labour Obfuscation in Sweatshop ProductionSweatshop garment factories represent a displacement AI relationship where skilled human labour of seamstresses/tailors is replaced by a system that obfuscates the often harmful working conditions and reduces human involvement to unseen, low-paid workers.
LLMs: Deskilling Human Essay Writing
Large Language Models (LLMs) are framed as capable of essay writing, but this relationship is a form of displacement. It deskills human users by reducing their role to mere 'prompt engineering,' obscuring the underlying exploited labour and hindering genuine learning.
- Displacement AI: LLMs replace human cognitive labour in essay writing.
- Deskilling Effect: Users' critical writing skills atrophy, replaced by superficial prompt interaction.
- Maximal Obfuscation: Hides 'ghost labour' of data annotators and sweatshop workers.
- Harmful Valence: Prevents users from learning 'anything substantial'.
AI Image Generation Workflow
AI image generators purport to create art, but this process displaces human artists and their creative labour. The workflow involves human prompts feeding into a system that relies on vast, often unethically sourced, datasets and hidden human-in-the-loop workers.
User Prompt (Input)
→
AI Image Generation (Processing)
→
Sweatshop Refinement (Hidden Labour)
→
Final Output (Often Plagiarized/Patchwork)
Chatbots & Illusion of Companionship
Harmful Valence of Chatbot CompanionshipChatbots designed for companionship displace genuine human social interaction, leading to potential psychological harm for users who mistakenly attribute minds or personhood to these inanimate objects. The labour of creating this illusion is maximally obfuscated.
Quantify Your Human-Centred AI ROI
Estimate the potential annual savings and hours reclaimed by strategically implementing human-centred AI in your enterprise.
Estimated Annual Savings
$0
Total Hours Reclaimed Annually
0
Your Human-Centred AI Implementation Roadmap
A phased approach to integrating AI that prioritizes human well-being, skill development, and ethical oversight.
Phase 1: Discovery & Strategy Alignment
Engage with stakeholders to understand existing workflows, identify cognitive labour points, and align AI strategy with human-centric values. This phase involves deep dives into current processes and potential areas for AI integration, focusing on enhancement over displacement.
Phase 2: Redefinition & Ethical Framework Development
Apply the proposed AI redefinition to classify potential AI projects as replacement, enhancement, or displacement. Develop a bespoke ethical framework that prioritizes human cognitive integrity, skill preservation, and transparency regarding human-in-the-loop dependencies.
Phase 3: Pilot Implementation & Sociotechnical Impact Assessment
Implement pilot AI systems with a focus on measurable human-centric outcomes. Conduct rigorous sociotechnical impact assessments to monitor effects on cognitive labour, skill development, and employee well-being. Adjust implementation based on feedback to mitigate deskilling and obfuscation.
Phase 4: Scaling & Continuous Human-in-the-Loop Optimization
Scale successful pilot programs across the enterprise, maintaining continuous oversight on human-in-the-loop involvement. Establish mechanisms for ongoing evaluation and adaptation, ensuring AI systems remain aligned with human values and do not inadvertently displace or devalue human cognitive contributions. Foster a culture of AI literacy and critical engagement.
Ready to Transform Your Enterprise with Human-Centred AI?
Don't let AI obscure human potential. Partner with us to design and implement AI solutions that truly empower your workforce and align with your ethical values.
Ready to Get Started?
Book Your Free Consultation.
Let's Discuss Your AI Strategy!
Lets Discuss Your Needs
Select a Date
Sun
Mon
Tue
Wed
Thu
Fri
Sat