Independent  ·  Responsible progress

The future is still a choice.

AI can cure disease, extend healthy lives, and lift human welfare, if we choose to build it responsibly.

We choose to keep building, with humility about the risks and determination to manage them. The question is not whether humanity should move forward, but how we can move forward well.

Our mission

Progress, with purpose.

Future of Choice is an independent organization dedicated to promoting the responsible development and use of artificial intelligence. We believe AI has immense potential to benefit society, from curing diseases and extending human lifespans to alleviating suffering and improving quality of life for all people.

Our mission is to offer an alternative perspective to the call for an immediate halt to AI research and development, which we believe is misguided and overlooks the positive outcomes AI can bring.

The case for building

The benefits are already becoming real.

This is not only a debate about distant possibilities. Carefully built AI is helping scientists, teachers and learners today, while opening new paths towards longer, healthier and more prosperous lives.

01

Medicine & health

In 2020, AlphaFold2 solved a 50-year-old problem: predicting a protein’s 3-D structure from its amino-acid sequence. Its predictions now cover virtually all ~200 million known proteins and are used by 2–3 million researchers in 190 countries, supporting work on disease, antibiotic resistance and drug discovery. Independent analysis links its use to a ~40% increase in submissions of novel experimental protein structures, and AlphaFold-linked research is about twice as likely to be cited in clinical articles.

02

Prosperity & work

AI can extend human capability, helping people learn, decide and create rather than simply removing them from the work. Used well, it can raise productivity and make valuable expertise more widely available. It works best when it supports people, enlarging what they can accomplish.

03

Knowledge & education

Personalised AI tutoring has produced measured maths gains in Ghana and English-language gains in Nigeria. The World Bank’s World Development Report 2026 also notes AI’s potential to ease constraints in health and education in developing countries, especially where access to skilled support is limited.

04

Science & the planet

AlphaFold is helping researchers design enzymes that can decompose plastics. Related research can advance materials, climate and biology work, turning better predictions into tools for understanding disease, protecting the living world and solving problems that have resisted decades of effort.

A practical view of safety

Progress and safety are not opposites.

AI deserves serious scrutiny. Its capabilities bring real risks, and good intentions alone are not enough. But safety is advanced by capable systems and serious research, not by stopping the work that helps us understand them.

Better evaluations can reveal weaknesses. Careful research can improve reliability and interpretability. Oversight, evaluation and transparency are how we make powerful systems worthy of trust.

The goal is not unchecked speed. It is steady progress with the knowledge, institutions and safeguards to guide it.

Oversight

Clear accountability, independent scrutiny and governance that can respond as capabilities change.

Evaluation

Rigorous testing, auditing and measurement before and after deployment, with room to learn from evidence.

Transparency

Cooperation among developers, independent experts and policymakers, with honest communication about limits.

Answering the concerns

Take the risks seriously.
Question the proposed remedy.

Calls to stop AI often begin from a sincere desire to protect people. We share that desire. The disagreement is about whether a halt would actually make anyone safer, and what we would give up along the way.

Concern 01

“A pause would keep us safe.”

A 2023 open letter gathered more than 30,000 signatures behind a six-month pause on training systems more powerful than GPT-4. But a voluntary moratorium is unenforceable and asymmetric: it binds mainly those already willing to follow the rules, while less careful actors can keep advancing.

Newer proposals, including a 2026 superintelligence ban built around a compute threshold, face the same problem. Capability spreads across borders and open weights, so a domestic halt mostly constrains those who already comply.

Concern 02

“Slowing down costs us nothing.”

Moving too quickly can cause harm, but delay has a real cost too. It is paid in suffering when better treatments, scientific tools and educational support arrive later than they could.

Every year a cure is postponed is a year people go without it. We should weigh the risks of action honestly, without treating stagnation as neutral.

Concern 03

“We must choose between safety and speed.”

This is a false dilemma. Safety is not a brake that only works when research stops; it is a discipline that must grow alongside capability.

We can fund technical safety research, build evaluation and audit regimes, and keep learning from the systems themselves. The answer is direction, not paralysis.

What we advocate

A future worth choosing.

We advocate the responsible development of AI, with its benefits shared broadly and its risks addressed directly. That means neither unchecked deployment nor an indiscriminate halt.

It means practical safeguards, honest evidence and cooperation across borders and institutions. We should remain ambitious about what AI can do for humanity, and equally ambitious about making it worthy of our trust.

  • Manage risks through proper oversight, collaboration and transparency.
  • Focus on creating safe, accurate, interpretable and trustworthy AI systems designed to serve humanity.
  • Foster cooperation between AI labs, independent experts and policymakers, and build robust governance and safety protocols.
  • Increase public funding for technical AI safety research.
  • Build institutions that help people through the economic and political disruption AI may cause.

Questions, answered

A clear position, without easy slogans.

Hope and caution belong in the same conversation. Here is where we stand.

Is Future of Choice against AI safety?

No. Safety is central to our mission. We support technical safety research, rigorous evaluation, independent oversight, transparency and robust governance. We believe these efforts become stronger when they develop alongside our understanding of real systems.

What exactly are you against?

We oppose calls for an immediate, indiscriminate halt to AI research and development. We do not oppose caution, targeted regulation or restrictions justified by evidence. Our concern is that a broad stop would sacrifice substantial benefits without reliably restraining the least careful actors.

Is a global pause even possible?

A voluntary pledge can change individual behaviour, and domestic law can constrain domestic developers. Neither creates a dependable global stop. Without worldwide enforcement, a pause mainly binds rule-followers, while capabilities continue to spread across borders and through open weights.

Will AI take everyone’s job?

AI will change work, and some economic disruption is likely. The honest response is neither to dismiss that uncertainty nor to assume the worst. We should design AI to augment people where possible, expand access to new capabilities and build institutions that help workers and communities adapt.

How can I help?

Read carefully, distinguish concrete evidence from speculation and share that evidence with others. Ask leaders to support technical safety research, independent evaluation and practical governance, while keeping cures, education and human welfare in the conversation. Responsible progress needs informed public support.