← All scenarios

Spaghetti Scenarios · Transcendence trajectory

Safe scientific and technological development

Science and technology advance without turning against those who create them. This includes the arrival of an artificial intelligence more capable than humans that stays aligned with our values: the same capability, without that alignment, becomes one of the greatest risks to humanity.

Science is already delivering results that can be verified. The rules for using artificial intelligence safely are only just starting.

Public investment in R&D

2.6%

of world output is invested in research and development (2023). This is public and private spending: there is no global series for public spending alone.

▲ Up from 2.1% in 2015

Alignment and audit frameworks

Since 2024

the European Union has an AI law in force; there are AI safety institutes in an international network and a UN scientific panel.

■ Progressing, but the high-risk rules were postponed to 2027–2028

Verifiable results in health and energy

260 M+

protein structures predicted by AI, open to everyone. In energy, laser fusion already delivers more than it receives.

▲ Fusion record in 2025: 4.13 times the laser energy

How much the world invests in research

Research and development spending as a percentage of world output. Move the slider to see each year.

0%1%2%3%19962005201520232.60%
  • R&D as % of world output2.60%in 2023

Results that can be verified

260 M+protein structures predicted by AI, in open access
2024 Nobel Prizein Chemistry, for predicting and designing proteins with computers
3.15 MJof fusion from 2.05 MJ of laser energy (2022): first ignition
×4.13record gain for laser fusion (2025)

Rules for safe artificial intelligence

This trajectory includes aligned superintelligence; its opposite, the same capability without alignment, is a global risk. These are the rules and institutions that have been created.

  1. 2023Bletchley Declaration (28 countries and the EU) and the first AI Safety Institute, in the United Kingdom.
  2. 2024The European Union's AI Act enters into force. The international network of AI safety institutes is born.
  3. 2025First International AI Safety Report (more than 100 experts). The UN creates its scientific panel on AI.
  4. 2026Second international report. The EU postpones its high-risk rules. The UN panel publishes a report on AI agents and loss of control.
  5. Today · October 2026
  6. 2027EU: high-risk rules for standalone AI systems (December 2).planned
  7. 2028EU: high-risk rules for AI built into products (August 2).planned

Safe scientific and technological development

Probability
Uncertain
Timeframe
Decades
Confidence
Low

Signals to watch

  • Public investment in R&D
  • Alignment and audit frameworks
  • Verifiable results in health and energy

Possible human action

  • Stable funding for basic science
  • Technology governance
  • Open scientific cooperation