World Solve

Governments currently discount long-term harms because future people cannot vote today. The open problem is creating formal institutions that represent future citizens in budget, infrastructure, climate, and technology decisions. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

As heat, flooding, drought, and erosion intensify, some communities will need to move. The challenge is designing relocation policy that protects land rights, livelihoods, identity, and consent rather than turning climate adaptation into forced displacement. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

Disasters, conflicts, and public-health crises force governments to buy quickly, often with reduced oversight. The open problem is building emergency procurement systems that are fast enough for crises but transparent enough to prevent leakage, favoritism, and fraud. The challenge is especially acute across fragmented municipalities, where legacy systems, coordination failures, and weak oversight can magnify harm.

Political communication will increasingly include deepfakes, voice cloning, and AI-generated persuasion at scale. The challenge is developing election rules, platform standards, and enforcement systems that protect speech while limiting manipulation. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

Future governments may depend on private firms for cloud services, AI tools, digital identity, and critical infrastructure software. The unresolved problem is designing procurement rules that preserve interoperability, portability, and long-term public control. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

As machine learning enters welfare, policing, housing, tax, and licensing systems, bias can become harder to detect and contest. The open problem is creating enforceable standards for fairness, appeals, and independent review across all public-sector uses of automated decision-making. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

Future states will need identity systems that work for migrants, displaced people, informal workers, and citizens missing formal paperwork. The challenge is building identity infrastructure that expands access without becoming a surveillance or exclusion tool. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

Governments will increasingly rely on automated systems to allocate benefits, flag risks, prioritize inspections, and recommend enforcement. The unresolved problem is creating governance structures that make these systems transparent, contestable, and accountable without disabling public service delivery. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

Patients waste time visiting multiple shops, while inventory data are private and substitution rules are complex. Privacy-preserving stock signals, route optimization, and pharmacist-led protocols could reduce search costs without creating a dominant intermediary. The opportunity is intentionally human-scale: a cooperative, clinic, workshop, school, or municipal team could pilot it within one locality and measure outcomes before expanding. Mathematical modeling can help with route uncertainty, incentives, or resource constraints; entrepreneurship can turn a reliable workflow into a viable service.

Automated trading algorithms can sometimes trigger rapid, extreme price swings in smaller or less liquid markets, especially during unusual conditions, without any single actor intending market harm. Building safeguards against this kind of systemic risk remains incomplete.

Networks of AI-assisted fake dating profiles can be deployed at scale to build trust with vulnerable users before extracting money or personal information. Detecting these coordinated networks before real harm occurs remains an unresolved platform challenge.

Safety filters designed to block clearly harmful requests can sometimes be bypassed through indirect or creatively phrased questions that still elicit dangerous information. Closing this gap while preserving the system's general usefulness remains an unresolved balancing act.

Counterfeit medications sold through unauthorized online channels can be difficult to distinguish from legitimate products without lab testing, putting patients at serious risk. Scalable, affordable detection at the point of purchase remains an unresolved challenge.

Malicious browser extensions can steal data or hijack sessions while appearing legitimate in official extension stores, and current vetting processes don't catch every harmful case before it reaches users. Strengthening this screening process remains an ongoing security challenge.

Despite improved detection tools, non-consensual intimate images can still spread quickly across platforms before removal, causing lasting harm to victims. Faster, more comprehensive detection and removal systems remain an urgent unmet need.

Content about topics like drug use, violence, or self-harm can be either genuinely educational or genuinely harmful depending on context and intent, and automated filters still struggle to reliably tell the difference. This creates ongoing tension between safety and access to legitimate information.

It's often unclear exactly what data went into training a given AI model, making it difficult to audit for bias, copyright issues, or harmful content. Greater transparency into training data remains technically and legally unresolved.

Automated moderation systems still struggle to tell the difference between genuinely harmful content and similar-looking legitimate content, like educational or satirical material. This tradeoff between over- and under-moderation remains unresolved at scale.

AI tools will increasingly support diagnosis and triage, but biases and opaque models may harm patients. Transparent validation frameworks, audit trails, and clinician–AI training curricula will be key.

AI tools will increasingly support diagnosis and triage, but biases and opaque models may harm patients. Transparent validation frameworks, audit trails, and clinician–AI training curricula will be key.