World Solve

AI and laboratory automation can identify promising compounds that fail during scale-up, processing, or deployment. The open problem is connecting discovery models to cost, supply, safety, and lifecycle constraints. The solution must work with limited capital, intermittent services, and local maintenance capacity. The research opportunity is to define measurable success criteria, test the approach in realistic settings, identify failure modes, and create a pathway from prototype to accountable deployment.

accelerating loss of the world's forests remains one of the most urgent challenges of our time. Systems currently in place are inadequate for the scale and complexity of the problem. A successful approach must work where resources are scarcest and demonstrate clear, measurable results. It should identify possible harms, respect local agency, and show how the solution can be maintained and expanded by the communities it is meant to serve.

Future progress on human development and planetary health depends on solving rapid death of coral reefs worldwide. Conventional approaches have proven insufficient. The required intervention must integrate community knowledge, data and appropriate technology while remaining affordable, privacy-preserving and operable with minimal external support. Strong solutions will articulate measurable impact metrics, risk mitigation strategies, and a plan for sustained operation by local actors after the pilot phase.

As digital navigation layers become permanent fixtures in daily life, physical landmarks lose meaning and people lose the ability to orient themselves without screens. The challenge is creating systems or practices that preserve spatial awareness and independent wayfinding while still allowing beneficial digital tools. Solutions must work across cultures, age groups, and infrastructure levels, avoid creating new dependencies, and remain usable when networks fail.

AI can generate convincing but fake expert testimony, reports, and analysis. The open problem is building public systems that preserve trust in real expertise while making fabricated authority easier to detect. The challenge is especially acute in informal settlements, where legacy systems, coordination failures, and weak oversight can magnify harm.

Future quantum networks will need coexistence with conventional fiber, satellites, and noisy metropolitan infrastructure. A major unresolved problem is developing practical repeaters, interfaces, standards, and operational security for real-world deployment. The solution must work with limited capital, intermittent services, and local maintenance capacity. The research opportunity is to define measurable success criteria, test the approach in realistic settings, identify failure modes, and create a pathway from prototype to accountable deployment.

The absence of robust, context-adapted solutions for institutional arrangements that accelerate the adoption of proven public health interventions creates cascading risks across interconnected human and natural systems. Key missing elements include durable performance, low lifetime cost, and governance arrangements capable of sustaining impact beyond pilot stages.

A persistent gap exists in realizing practical, scalable solutions for methods for restoring the carbon and biodiversity values of degraded peatlands and wetlands. Scientific understanding, engineering readiness and institutional capacity remain insufficient to translate promising concepts into durable impact under real-world constraints of cost, skills and governance.

A persistent gap exists in realizing practical, scalable solutions for technologies for detecting and neutralizing chemical and biological threats at public venues. Scientific understanding, engineering readiness and institutional capacity remain insufficient to translate promising concepts into durable impact under real-world constraints of cost, skills and governance.

Verifying that voting machine software hasn't been secretly altered requires specialized technical audits that aren't consistently performed everywhere elections take place. Making this verification standard and accessible remains an unresolved election security challenge.

Employee behavior remains one of the most common causes of security breaches, yet many small businesses lack the resources for regular, effective cybersecurity training. Scaling affordable, effective training globally remains an unresolved practical challenge.

As detection tools improve, so do the techniques used to generate reviews designed specifically to slip past them, creating an ongoing arms race. Staying consistently ahead of these evasion tactics remains technically unresolved.

Recommendation systems tend to show users more of what they already agree with because that content performs better for engagement, gradually narrowing exposure to differing viewpoints. Correcting for this without reducing relevance remains an unresolved design challenge.

Blockchain can record ownership transactions, but it doesn't inherently verify that the underlying digital artwork itself is original or hasn't been duplicated elsewhere. Solving this deeper authenticity verification remains an unresolved technical gap.

Insurers offering cybersecurity coverage often struggle to accurately assess a company's true risk level, leading to mispriced policies that either overcharge safe companies or undercharge vulnerable ones. Better, more accurate risk assessment tools remain underdeveloped.

Techniques that let multiple parties jointly compute something without revealing their private data to each other exist, but they remain too slow and complex for widespread practical adoption. Making this both secure and fast enough remains an active research challenge.

AI-generated contracts, court filings, or legal notices can appear authentic enough to deceive people unfamiliar with proper legal formatting, creating new fraud risks. Detecting these fakes reliably remains an emerging unresolved challenge.

AI models trained on real-world data often absorb and sometimes amplify existing societal biases, and current bias correction methods only partially address the problem. A more complete solution remains an active and unresolved area of research.

Humans can learn a new task from a handful of examples, while most AI systems still need massive amounts of training data to perform reliably. Closing this efficiency gap remains one of AI research's central unsolved challenges.

Robots can now handle many industrial tasks well, but tasks requiring soft, adaptable touch, like handling fabric or fragile objects, remain surprisingly difficult. Closing this dexterity gap would open robotics to far more everyday uses.