AI-directed research produces a reproducible discovery step

Technology and infrastructureUpside scenario

Independent laboratories verify that AI-directed research systems can formulate hypotheses, run automated experiments and produce useful, reproducible results in three commercially important fields: therapeutic candidates, battery and catalyst materials and engineered enzymes. Licensed and open systems become available within a year. For qualifying programmes, the time from target to validated candidate halves and cost per validated candidate falls 50 per cent over three years. General superintelligence is not assumed at Standard Severe; the first-pass takeoff is carried as the Extreme setting. Nations must supply laboratories, compute, skilled evaluators and approval pathways to capture discoveries, while managing dual-use access and concentration of intellectual property.

Most Relevant Trends

The 14 trends graded High or above, of the 82 of 250 relevant to this scenario: those that change how likely it is, how hard it lands, or are themselves sharply changed by it. Each is graded on a seven-point scale from Nil to Critical, with the mechanism the grade rests on.

By capability domain

The grades of each domain’s relevant trends, added together.

Critical Technology

Production & Innovation

Trade & Investment

Human Capital

Governance & Integrity

National Security

Strategic Infrastructure

Financial Strength

Information & Influence

By relevance

Critical

Very high

High

Moderate

Low

Very low

1 trend

6 trends

7 trends

20 trends

24 trends

Trend	Relevance	Why
AI for science Critical Technology	Critical	The scenario is the culmination of AI-accelerated science in therapeutics, materials and enzymes.
Frontier AI capability Critical Technology	Very high	Autonomous reasoning and task execution are the core capability that makes AI-directed research systems possible.
AI agents Critical Technology	Very high	Autonomous reasoning and task execution are the core capability that makes AI-directed research systems possible.
AI regulation & safety Governance & Integrity	Very high	Enforceable evaluation and dual-use controls decide whether licensed and open research systems can be deployed safely.
AI job displacement Human Capital	Very high	Growing agentic AI capability underlies AI-directed research and the skilled evaluators needed to capture it.
R&D intensity Production & Innovation	Very high	Capturing AI-directed discoveries requires laboratories, skilled evaluators and pathways that research spending builds.
AI drug discovery Production & Innovation	Very high	Therapeutic candidates are a named field; nations with health data, compute and validation labs capture the step.
Biosecurity & dual-use biology National Security	High	AI-directed design of enzymes and therapeutics widens dual-use access, making screening and laboratory oversight central.
Gene & cell therapies Critical Technology	High	Gene editors are engineered enzymes; specialist manufacturing and clinical capacity decide whether AI-designed candidates become national treatments.
Skilled talent migration Human Capital	High	Clustering of research talent largely determines which nations can supply evaluators and capture discoveries.
Automated AI research Critical Technology	High	Automated experimentation with evaluation is the precursor of AI-directed science and drives the Extreme takeoff setting.
Synthetic biology Critical Technology	High	Engineered enzymes and therapeutics are named fields; biofoundries supply the automated experiments and scale-up.
Drug pricing pressure Production & Innovation	High	Halved cost per validated candidate is the answer to research budgets squeezed by price controls.
Market concentration Production & Innovation	High	Compute and model owners capture IP from AI discovery; the scenario names concentration as a management test.

Top 5 Scenario Trends – Regional Attention

For the 5 most relevant trends: the regions, blocs and countries where each draws the highest attention intensity, with its rank among the 250 trends there.

AI for scienceCritical