Open scientific research

Scientific results should be available to everyone as a strong default, because publicly relevant knowledge serves humanity best when shared and most science builds on open prior work; but the default is rightly bounded where results are genuinely dangerous, legitimately owned, or meaningless without the capacity to use them.

• Results should be open by default because science is cumulative and openness accelerates discovery.
  Why Researchers build on each other's findings, so locking results behind paywalls or secrecy slows the collective process and duplicates effort, which means openness is not charity but the engine of progress.
  Example The rapid global development of COVID-19 vaccines depended on the open sharing of the viral genome sequence, which let labs worldwide start work simultaneously (from_etg_textbook.md, as_of 2026-06).
  Then evaluate But 'accelerates discovery' is an argument for sharing among researchers, which does not yet settle whether every result should reach literally everyone.
• Publicly funded results have a particular claim to openness, because the public already paid for them.
  Why When taxpayers fund research, the findings are in a real sense already owned by the public, so charging the same public again to read them is a double payment that open-access mandates exist to end.
  Example The global open-access movement and funder mandates requiring publicly funded research to be freely available rest on exactly this ownership logic (global pattern, as_of 2026-06).
  Then evaluate Yet this argues from who paid, so it weakens for privately financed research, where the funder has a defensible claim to the results.
• The default breaks where results are dangerous, because 'everyone' includes those who would weaponise them.
  Why Some findings are dual-use, the same method that cures can also harm, so publishing the recipe to everyone hands capability to malicious actors, which is a reason to restrict not the finding but the actionable detail.
  Example Debates over publishing gain-of-function research that could make pathogens more transmissible show why some results are deliberately withheld or restricted (global_evidence.md, as_of 2026-06).
  Then evaluate The hinge: openness about 'what was found' can be safe while openness about 'how to do it' is not, so the limit is on dangerous detail, not on knowledge as such.
• Openness is also hollow where the audience lacks the capacity to use or interpret the results.
  Why Raw data and technical findings released to a public without the training to read them can mislead as easily as inform, so 'available to everyone' requires translation and context, not just publication.
  Example The pandemic showed how openly available but misread research fuelled misinformation, with falsehoods like Singapore's HardwareZone fake COVID-death post moving real behaviour faster than corrections (singapore_evidence.md, as_of 2026-06).
  Then evaluate So availability is necessary but not sufficient: results dumped on a public without interpretation can do harm openness was meant to prevent.

The statement conflates three different things, the knowledge, the application and the ownership, and is right about one and wrong about the others: the knowledge should indeed be open to everyone, but the dangerous application should not, and the legitimately owned application cannot be, so the answer depends entirely on which of the three 'results' means.

• In the domain of knowledge, results should be open to everyone, with almost no exceptions.
  Why Understanding how the world works is a public good that costs nothing to share and harms no one by being known, so the basic findings of science belong to humanity by their nature.
  Example The open publication of the COVID-19 genome let the whole world understand the threat at once, knowledge that endangered no one by being shared (from_etg_textbook.md, as_of 2026-06).
  Then evaluate So for pure knowledge the statement is simply correct, which is why the interesting cases lie in the other two domains.
• In the domain of dangerous application, results should not be open to everyone, because capability is not the same as understanding.
  Why A method that lets someone build a weapon or enhance a pathogen is actionable in a way pure knowledge is not, so making it available to everyone, including bad actors, creates risk the underlying knowledge does not.
  Example Restrictions on publishing the operational details of dual-use research, such as making a virus more transmissible, withhold the application while the science remains understood (global_evidence.md, as_of 2026-06).
  Then evaluate The complication for the statement: here 'available to everyone' is not a virtue but a hazard, so the same word that was right for knowledge is wrong for application.
• In the domain of ownership, results often cannot be open to everyone, because private investment created them on the promise of return.
  Why When firms fund research expecting patents or profit, removing their exclusivity would destroy the incentive to fund it at all, so forced openness can mean less research gets done.
  Example The vaccine-patent debate during COVID pitted the case for sharing recipes against pharmaceutical firms' claim that exclusivity funds the next breakthrough, a genuine clash of openness and incentive (from_etg_textbook.md, as_of 2026-06).
  Then evaluate Yet ownership is a policy choice, not a law of nature: society can shorten patents or fund research publicly, so this limit is negotiable in a way the danger limit is not.
• Keeping the three domains separate yields a usable rule the blanket statement cannot.
  Why If knowledge is open, dangerous application is restricted, and owned application is negotiated, then each kind of result gets the treatment its nature requires, which a single 'available to everyone' cannot deliver.
  Example The coexistence of open genome data, restricted dual-use methods and patented drugs shows science already operating on exactly this three-way distinction (global pattern, as_of 2026-06).
  Then evaluate The reframe's payoff: 'how far do you agree' has no single answer because 'results' is not one thing, so the right reply specifies the domain.