Radical Reflections

      Tech and futures blog | Where ideas in AI, design, human cognition, and futures converge. Thinking out loud — in pursuit of what matters next.

      Tech and futures blog | Where ideas in AI, design, human cognition, and futures converge. Thinking out loud — in pursuit of what matters next.

      Generative AI and the SDGs: The Goals Were Built for One World. We’re Already Living in Another.

      As generative AI moves from tool to infrastructure, the SDGs must adapt — or risk becoming well-intentioned relics.

      Thought Exploration Series

      Rabih Ibrahim

      Rabih Ibrahim

      10 min read
      August 14, 2025

      2015, the United Nations adopted the Sustainable Development Goals — 17 global priorities to guide humanity toward a more just, equitable, and sustainable future by 2030. It was a rare moment of collective clarity. Governments, institutions, civil society, and the private sector agreed — at least in principle — on what progress should look like.

      But the SDGs were designed for a world governed by human systems. A world where policies were debated, where institutions held the levers of implementation, where technology was a tool — not a co-author.

      We don’t live in that world anymore.

      Generative AI isn’t just speeding up development. It’s changing the conditions under which development occurs. It’s altering who gets to define a solution, whose knowledge is represented, and which values are embedded in the systems we build. And while the 2030 goals remain relevant, their foundation is under strain.

      . . . . .

      From Human Systems to Hybrid Systems

      The SDGs were built on an assumption: that development would be led by people — by ministries, NGOs, educators, scientists, communities. The institutions might have been flawed, but at least they were human.

      Now, the terrain is shifting. AI tutors are reaching students before ministries of education can update textbooks. Health tools are generating diagnoses ahead of local policy review. Generative models simulate climate scenarios before environmental regulations are drafted. In politics, even heads of government are experimenting: Sweden’s prime minister, Ulf Kristersson, said he uses ChatGPT and France’s LeChat for a “second opinion,” sparking backlash — “we didn’t vote for ChatGPT” — and igniting a wider debate about AI in decision-making. The very systems that were once the gatekeepers of change now find themselves reacting to tools that move faster than legislation, funding cycles, or public consensus.

      It’s not just a matter of scale. It’s a shift in agency. Who Defines Progress?

      . . . . .

      Who Defines Progress?

      Generative systems don’t just accelerate action — they redefine the terms of action. A chatbot might offer education, but what values does it teach? An AI-generated climate brief might be technically correct, but who’s prioritized in its assumptions? Efficiency doesn’t equal equity. Speed doesn’t mean justice.

      This is not a rejection of AI’s potential. It’s a reckoning with the fact that development is no longer a purely human project. When tools begin to shape not only the process of progress, but the vision of what progress looks like, we need a new lens. And perhaps, new goals.

      The question becomes: can we pre-design this new factor — not just to solve problems faster, but to help us think and act as one? And by “us,” I mean more than humanity. I mean the citizens of the world in the widest sense: people, nature, other creatures whose futures are bound to ours.

      . . . . .

      Goal-by-Goal: How GenAI Is Being Used to Meet — Or Complicate — the SDGs

      Let’s look at how generative AI is already intersecting with the SDGs — not in theory, but in practice.

      SDG 1: No Poverty

      Expanding access — but risking deeper exclusion

      • HelloTask and Harvesting use AI to connect domestic workers and smallholder farmers to microloans and income tracking tools.
      • Leaf Global deploys chatbots for financial inclusion among refugees, offering financial literacy and transaction capabilities in low-connectivity settings.
      • Mastercard’s Inclusive ID pilot uses AI to help build digital financial identities for people without formal documentation.

      . . . . .

      SDG 2: Zero Hunger

      Optimizing food systems — but not political access

      • PlantVillage Nuru uses AI-powered image recognition to identify crop diseases directly from farmers’ smartphones.
      • OneSoil and IBM’s PAIRS apply AI to simulate crop yields and weather patterns for improved planting strategies.
      • Models like ProGen are being used to accelerate the design of synthetic proteins for alternative food sources.

      . . . . .

      SDG 3: Good Health and Well-being

      Enhancing diagnostics — without systemic safeguards

      • Google’s Med-PaLM and Glass AI provide GenAI-driven clinical decision support for medical practitioners.
      • Mental health chatbots like Woebot and Wysa deliver evidence-based cognitive behavioral interventions through conversational AI, expanding access in regions with few therapists.
      • AI-driven maternal health programs, such as SuperHumanRace, use predictive models to identify pregnancy-related risks early.

      . . . . .

      SDG 4: Quality Education

      Personalized instruction — but epistemic narrowing

      • Khanmigo and Squirrel AI adapt lesson plans and pacing to individual students’ progress.
      • Class Companion uses large language models to simulate interactive writing feedback for learners.
      • UNESCO promotes AI-assisted generation of multilingual curricula to reach linguistically diverse classrooms.

      . . . . .

      SDG 5: Gender Equality

      Bias detection — within biased systems

      • Tools like Textio and Applied use AI to detect and correct gender bias in hiring processes and job descriptions.
      • AI supports maternal health diagnostics by analyzing imaging and patient data for early-stage complications.
      • In mental health, AI-powered platforms track and analyze reports of gender-based violence, helping NGOs identify patterns and direct interventions more effectively.

      . . . . .

      SDG 6: Clean Water and Sanitation

      Sensing contamination — still failing enforcement

      • IBM Green Horizons uses AI to monitor water quality and model contamination risks.
      • Microsoft’s AI for Earth program funds AI-based flood prediction and drought forecasting tools.
      • Satellite imagery analysis with machine learning enables NGOs to detect illegal water diversion.

      . . . . .

      SDG 7: Affordable and Clean Energy

      Energy efficiency — offset by AI’s own appetite

      • DeepMind optimized Google’s data center cooling systems, reducing energy use by 40%.
      • Microsoft’s Net Zero Copilot helps organizations model emissions reduction scenarios.
      • Grid optimization systems integrate AI to balance renewable energy sources with fluctuating demand.

      . . . . .

      SDG 8: Decent Work and Economic Growth

      Augmenting tasks — displacing roles

      • GenAI is embedded in productivity tools like Jasper, Writer, and Slack GPT to automate repetitive tasks.
      • Accenture projects 31–47% of current work hours could be automated by AI by 2030.
      • Pymetrics is a talent-matching platform that uses neuroscience-based games and AI analysis to assess candidates’ cognitive, emotional, and social traits, matching them to roles where they are most likely to succeed.

      . . . . .

      SDG 9: Industry, Innovation, and Infrastructure

      Accelerating R&D — but centralizing control

      • AI accelerates materials science and drug discovery via platforms like Insilico Medicine and Atomwise.
      • Generative design tools, such as Autodesk’s platform, produce optimized prototypes for manufacturing.
      • Infrastructure resilience modeling uses AI to simulate how systems respond to extreme weather and other stresses.

      . . . . .

      SDG 10: Reduced Inequalities

      Scaling access — without closing the gap

      • GenAI supports real-time translation and voice interfaces, improving accessibility for non-literate and non-native speakers.
      • Platforms like Talk2U engage youth in underserved urban areas through AI-driven educational content.
      • The Inclusive Internet Index, supported by Meta and the Economist Intelligence Unit, uses AI-driven data analysis to identify digital access gaps and inform targeted connectivity projects in low-income regions.

      . . . . .

      SDG 11: Sustainable Cities and Communities

      Designing urban futures — with contested authorship

      • Sidewalk Labs’ Delve uses AI to generate neighborhood plans balancing factors like sunlight, density, and amenities.
      • GenAI simulations model climate-resilient infrastructure designs for urban settings.
      • AI-driven digital twins of cities track emissions, waste patterns, and mobility flows.

      . . . . .

      SDG 12: Responsible Consumption and Production

      Synthetic sustainability — real-world extraction

      • AI systems track materials and supply chains for sustainability reporting.
      • Retailers use AI to forecast demand and reduce overstocking, minimizing waste.
      • SAP’s GreenToken platform applies blockchain and AI to verify ethical sourcing.

      . . . . .

      SDG 13: Climate Action

      Better models — at a hidden cost

      • ClimateGPT prototypes simulate climate scenarios and test potential policy interventions.
      • AI helps NGOs detect illegal logging, emissions fraud, and deforestation via satellite imagery and automated pattern recognition.
      • The Climate TRACE coalition uses AI and remote sensing to independently track greenhouse gas emissions from over 70,000 sites worldwide, providing public, verifiable climate data.

      . . . . .

      SDG 14: Life Below Water

      Monitoring oceans — without changing behaviors

      • AI systems classify coral reef health using underwater image datasets.
      • Acoustic monitoring combined with AI detects illegal fishing activities.
      • Machine learning models track marine pollution movement via satellite imagery.

      . . . . .

      SDG 15: Life on Land

      Forest awareness — without forest protection

      • Global Forest Watch uses AI to identify deforestation in near-real-time.
      • GenAI generates environmental impact reports for use by policymakers and conservationists.
      • Audio AI tools monitor protected areas for sounds associated with poaching.

      . . . . .

      SDG 16: Peace, Justice, and Strong Institutions

      Accelerating disinformation — eroding trust

      • AI-powered forensic tools analyze and authenticate digital media for use in legal proceedings.
      • Generative detection systems identify deepfakes and manipulated content at scale.
      • Machine learning platforms track misinformation networks and their spread across social platforms.

      . . . . .

      SDG 17: Partnerships for the Goals

      Tech platforms as partners — with opaque incentives

      • AI is becoming embedded in the data and coordination systems that underpin global collaboration.
      • The UN’s Global Digital Compact and Digital Public Infrastructure initiatives now integrate AI platforms for real-time SDG progress tracking and open data sharing.
      • Partnerships with Microsoft, Google, and IBM have produced AI tools for mapping climate risks, translating UN resources into multiple languages, and improving humanitarian logistics.
      • Open-source AI projects, such as those supported by the Global Partnership on AI, are developing multilingual chatbots and data analysis tools to support NGOs and community-level SDG reporting.

      . . . . .

      What’s Actually Happening Here?

      These tools aren’t simply speeding up progress — they’re expanding the ways it can happen. In the process, they’re also reshaping what “progress” looks like, sometimes in ways we don’t fully anticipate. They bring new capacities, but also embed the assumptions and priorities of their makers. They can create pathways we didn’t have before, and dependencies we haven’t fully considered. And they are shaping, in real time, the very realities and futures the SDGs were designed to address.

      The question is no longer just whether GenAI aligns with the SDGs.
      It’s whether the SDGs can evolve quickly enough to guide a world that is learning, generating, and adapting in real time.

      . . . . .

      Beyond Metrics: Rethinking the Frame

      The SDGs still matter. They offer a moral compass — a shared map of long-term priorities like health, equity, climate, and dignity. But that map was drawn for a different terrain. A slower world. One where change followed deliberation, where governance had the upper hand, and where development was driven by institutions, not algorithms.

      Today, the landscape is being redrawn by generative systems that don’t wait for consensus. They iterate, scale, and embed themselves — shaping decisions, redistributing agency, and sometimes rewriting the rules entirely. The development process is no longer solely in human hands.

      Perhaps we need to think of the SDGs less as fixed endpoints and more as vectors, in the sense Kevin Kelly describes in The Inevitable — forces that set direction even as the terrain shifts beneath us. This perspective would allow them to guide momentum, not just measure milestones, making them relevant in a landscape where change is constant and non-linear.

      That doesn’t make the SDGs obsolete, but it does make them fragile. Static goals can’t guide dynamic systems. What we need now are frameworks that aren’t just measurable, but adaptable — capable of revising themselves without collapsing. To do this, they must be embedded into the evolving architectures of governance, technology, and civic participation, so they can remain present and influential even as the systems around them transform.

      Maybe it’s time to ask: should there be an SDG 18 — Responsible AI and Digital Governance? A goal that acknowledges the infrastructural role of intelligence systems, and centers questions of accountability, access, and ethical design. Not just as an afterthought to development, but as one of its foundations. And perhaps, in an AI-mediated world, this becomes the goal that matters most — because without it, the rest may never be realized as intended.What Kind of Goals Survive a Self-Updating World?

      . . . . .

      What kind of Goals Survive a Self-Updating World?

      The SDGs gave us a moment of rare global alignment. But the world they were designed for is already behind us. The one we’re moving into demands new strategies — not just to implement policy, but to co-evolve with systems that learn, simulate, and act on their own timelines.

      What kind of goals can survive in a world that continuously rewrites its own rules?
      Can we design for long-term equity in environments optimized for short-term efficiency?
      And if generative AI becomes part of the invisible infrastructure of development — shaping decisions, rewriting knowledge, mediating access — who do we hold accountable when the future accelerates without consent?

      We may need more than new technologies to answer these questions.
      We may need a deeper political imagination.

      . . . . .

      Questions We Can’t Ignore

      • What kind of goals can survive in a world that updates itself faster than our policies?
      • Can long-term equity exist in systems optimized for short-term reward?
      • And if AI becomes part of the infrastructure of development — who do we hold accountable when the future accelerates without consent.

      . . . . .

      References

      . . . . .