The most ambitious form of ecosystem leadership is not competing in the market. It is defining the playing field itself. When you help develop industry standards and best practices, you are setting the rules everyone operates within, establishing what good looks like, and preventing harm across an entire industry rather than one business. Standards sound dull next to thought leadership or community building, and that is precisely why the influence available there is so poorly contested. Help establish a standard that gets adopted, and your thinking shapes how thousands of organizations work.

Why Standards Matter: The Invisible Infrastructure of Progress

Most people do not think about standards until they are absent, and then they notice immediately. Consider what came before electrical standards: every town ran different voltage and frequency, and factories had to be custom-built for each location they opened in. Once standards emerged, electricity could be generated anywhere and used anywhere, and the standard unlocked an enormous amount of economic value that had previously been consumed by adaptation. Nobody experienced that as a new invention. They experienced it as friction disappearing.

The same logic applies to AI. Without standards for data privacy, every company invents its own security protocols. Without standards for model evaluation, every vendor claims superiority using metrics of its own choosing, which makes comparison impossible and buying decisions arbitrary. Without standards for ethical AI development, companies build safety practices inconsistently, and the practices that exist depend on who happened to be in the room. The absence of standards produces inefficiency, duplicated effort, and consumer confusion, all of which are paid for by somebody.

Standards deliver five specific benefits. Safety and risk reduction: they establish minimum requirements that prevent harm. Quality assurance: they establish benchmarks so that products can be compared fairly rather than rhetorically. Efficiency: companies stop solving identical problems independently. Market development: industries with strong standards attract more investment and talent because risk is lower and therefore easier to price. Innovation enablement: by establishing baseline practices, standards free companies to innovate above the baseline instead of rebuilding the foundation every time.

The Paradox of Standards and Innovation

Many people fear that standards restrict innovation. The opposite is closer to the truth. Without standards, companies waste resources reinventing foundations that nobody differentiates on anyway. With standards, they can innovate rapidly on top of established practice. The internet would not exist without TCP/IP standards. Modern software would not exist without programming language standards. Standards and innovation are not opposed forces to be balanced against each other; innovation tends to accelerate once the standards beneath it are settled.

Types of Standards and Where They Come From

Not all standards are created equally, and understanding where a given standard comes from is how you work out where your own contribution would count for something. The four sources differ in how long they take to produce, how much authority they carry, and how easy they are to influence from outside, and those three properties usually move in opposite directions to each other.

Formal Standards

These are developed by official standards bodies such as ISO, IEEE, and government agencies, through rigorous consensus-based processes. They typically take years to develop but carry significant authority once published. Examples include ISO 27001 for information security, IEEE standards work covering AI systems, and government AI regulations still being written. Formal standards are valuable but slow, which makes them best suited to mature areas where broad consensus already exists. In a rapidly evolving field like AI, formal standards lag the technology by construction.

Industry Guidelines

These are developed by industry consortia, professional associations, and non-profit organizations. Examples include the Partnership on AI principles, IEEE Ethically Aligned Design, and the enterprise practice guidance published by major technology companies. Guidelines are faster to develop and more flexible than formal standards, and the tradeoff is exactly what you would expect: they carry less mandatory weight but often achieve more practical adoption, because organizations can implement them without waiting for a committee cycle to complete.

De Facto Standards

These emerge from market adoption rather than formal process. JSON became the default for data interchange because everyone adopted it, not because a committee mandated it. Linux became a de facto standard for server infrastructure the same way. De facto standards emerge when one approach proves so much more workable than the alternatives that adoption becomes inevitable. They are powerful, and they carry a specific risk: once an approach achieves dominance it can lock in choices that were never optimal, simply because the cost of moving has become collective.

Company-Specific Standards

Large companies establish practices that others follow. When Apple sets a privacy standard in its products, competitors feel pressure to match it. When leading AI companies publish responsible AI frameworks, others adopt similar approaches rather than defend the absence of one. Company-specific standards influence the market even when they are never formally stated as standards at all, which means an organization with a strong reputation can set a de facto standard through its own practices long before any body writes it down.

Emerging Standards in AI

The AI standards landscape in 2026 is still forming, which is unusual and temporary. Most fields present a settled map that newcomers learn; this one presents open questions across six areas at once, each of which will be answered by whoever does the work. Leading on standards in any of these areas positions you at the forefront of industry evolution, and the barrier to entry is currently lower than it will ever be again.

Contributing to Standards Development: A Practical Path

Standards development is not closed to outsiders, which surprises people who assume the work happens somewhere they would need an invitation to reach. It requires persistence and substantive contribution, but practitioners at every level can influence standards, and the path runs through six steps in a deliberate order. Skipping to the end, which usually means proposing a standard before establishing that a gap exists, is the most common way well-intentioned contributions get ignored.

1. Identify the Standards Gap

Start by identifying what is missing. Are there critical practices that have not been formalized? Questions the industry answers inconsistently? Problems you see recurring across organizations rather than only in yours? Document the gap, write about it, and collect evidence from your own experience to show that the problem is real and affects multiple organizations rather than being a local irritation. Two examples of a well-stated gap: every company invents its own approach to documenting model limitations, so a documentation standard would improve transparency; and different organizations evaluate AI fairness using different metrics, which makes comparison impossible.

2. Research What Exists

Before proposing anything new, understand what is already being done. Are there standards bodies working on this? Industry groups? Companies with established practices? Academic research? Build on existing work rather than creating from scratch, both because duplication wastes your effort and because arriving with an obvious ignorance of prior work costs credibility you will need later. This research also tells you where to contribute: if a working group is already developing standards in your area, that is where to engage, and if no formal work exists, you may need to initiate a community effort.

3. Engage in Existing Standards Bodies

Do not wait to be invited. Most standards bodies welcome participation, so join committees, attend working groups, and volunteer. What earns you standing there is substantive work: research, case studies, implementation experience, and detailed feedback on drafts. Find where the work is actually happening, whether at ISO, IEEE, professional associations, or consortia, and participate consistently rather than episodically. This is the mechanism by which emerging standards get influenced, and it is available to anyone willing to do the reading.

Where the AI Standards Work Happens

The landscape divides into five categories. Government-led work sits with national standardization institutes and regulatory bodies, including the frameworks built around the EU AI Act. International work runs through the ISO and IEC joint technical committee structure that handles AI. Non-profit work includes IEEE, the Partnership on AI, and the Responsible AI Institute. Industry consortia cover sector-specific working groups and major technology company initiatives. Professional associations include the ACM and domain-specific bodies. Identify which of these are most relevant to your focus area and engage there rather than everywhere.

4. Publish Case Studies and Evidence

Standards are most credible when grounded in evidence rather than argument. Document how you have addressed the problem you want standardized, publish case studies showing what worked and what did not, and share the data from your experience including the parts that complicate your position. An example of the form: we implemented a fairness evaluation process for our healthcare AI models, and here is how we did it, the challenges we hit, and the results. A case study of that kind becomes evidence that informs standards development in a way that opinion cannot.

5. Build Coalitions

Significant standards emerge when multiple organizations agree, not when one organization is right. Build coalitions of like-minded leaders who support the standard you are developing, collaborate on research, co-author papers, and present findings together rather than separately. Coalitions carry more weight than individuals for a simple structural reason: when ten respected organizations publicly commit to a standard, the market reads that as a shift in what is expected, whereas a single organization advocating a standard reads as a company describing its own practice.

6. Start With Best Practices, Evolve to Standards

Do not jump straight to formal standards development. Start by documenting best practices in the form "here is what works, based on our experience", publish those, and refine them based on the feedback you get. Once best practices are well documented and widely adopted, they are ready to formalize into standards. This evolutionary path, running from experience to best practices to guidelines to formal standards, works better than trying to skip steps, because each stage generates the evidence and the constituency the next stage requires.

Building Best Practices Within Your Organization

You do not have to change industry-wide standards in order to establish standards. Often the most powerful first step is establishing them within your own organization or industry segment, where you have the authority to simply decide and the visibility to see whether the decision works. That local record is also what makes any later contribution credible, since standards bodies weigh implementation experience far more heavily than proposals.

Create documentation covering how you evaluate AI models for fairness and accuracy, how you document model limitations and uncertainty, how you implement responsible AI practices, how you conduct security testing, how you manage training data ethically, and how you monitor deployed models for performance degradation. Then share those standards with your community and make them the expected baseline in the conversations you are part of. When others see your standards working in practice, they adopt them, and over enough time and enough organizations, best practices become industry norms.

The Politics of Standards Development

Standards development is partly technical and partly political, and understanding the political half is what lets you navigate it without becoming cynical about it. Competing interests come first. Different organizations want different standards because different standards favour their existing strengths. The organization with strong privacy practices wants privacy standards. The organization with strong performance wants performance standards. Neither position is dishonest, and standards development requires negotiating between them rather than pretending they do not exist.

Power dynamics come second. Large companies hold outsized influence because they have the resources to contribute sustained attention, which is the actual currency of committee work. That is a reason to actively ensure diverse voices participate rather than a reason to withdraw, because minority viewpoints strengthen standards by forcing them to address edge cases and needs the dominant participants do not encounter. A standard written entirely by the well-resourced will be perfectly workable for the well-resourced.

Adoption challenges come third. Standards only matter if they are adopted, and the most technically perfect standard that nobody uses is useless. Practical standards that companies can realistically implement spread faster than idealistic standards that are expensive to comply with. Evolution over time comes fourth. Good standards evolve: document your assumptions, plan for regular review, and be willing to update as technology changes, because standards that cannot evolve stop being enablers and become the obstacle everybody works around.

Making Standards Practical and Adoptable

Effective standards share six properties. They are specific enough to be actionable, flexible enough to work across contexts, evidence-based rather than asserted, inclusive of diverse perspectives, achievable for practitioners at varying resource levels, and reviewable so that they can evolve. The resource point is the one most often missed. Standards that favour only large, well-resourced organizations do not last long, because the organizations excluded by them route around them and the standard fragments. The best standards scale from startups to enterprises.

The Leverage of Standards Leadership

Leading on standards development feels less visible than thought leadership or community building, and the influence is considerably larger. When you help establish a standard the industry adopts, you have shaped behaviour across thousands of organizations, most of which will never know your name and all of which will operate differently because of work you did. That is a form of impact that outlasts the roles and companies it was produced from.

Standards leaders also earn specific benefits. Deep credibility, because you are not just speaking or advising, you are shaping how the industry operates. Relationships, because standards work connects you with decision-makers across many organizations at once, in a working context rather than a networking one. First-mover advantage, because organizations that implement your standards before wide adoption are already compliant when everyone else starts. And influence without authority, since you do not need a title to influence standards; substantive contribution carries the weight on its own.

Anti-Patterns to Avoid

• Proposing a standard before documenting the gap. Without evidence that the problem is real and recurs across organizations, a proposal reads as a preference.
• Skipping the research step. Arriving with no knowledge of existing work duplicates effort and costs the credibility you will need later.
• Waiting to be invited. Most standards bodies welcome participation, and the people shaping outcomes are the ones who showed up.
• Contributing opinions rather than work. Standards are influenced by those who do the research, the implementation, and the evidence collection.
• Jumping straight to formal standardization. Skipping the path from experience to best practices to guidelines removes the evidence and the constituency the next stage needs.
• Advocating alone. A single organization promoting a standard reads as a company describing its own practice.
• Designing for well-resourced organizations only. Standards that smaller practitioners cannot afford to implement fragment rather than spread.
• Optimizing for technical perfection over adoptability. The most rigorous standard that nobody implements changes nothing.
• Treating standards as finished. Undocumented assumptions and no review cycle turn an enabler into an obstacle as technology moves.
• Pretending the politics are not there. Competing interests are structural, not a sign of bad faith, and they have to be negotiated rather than ignored.
• Assuming you need industry scale to start. Standards established inside one organization are both real and the credential for everything after.

Practice Prompts

• Name the gap. "Based on these recurring problems I have seen across organizations, draft a clear statement of a standards gap: what is inconsistently done today, why that inconsistency causes harm or waste, and what a standard would need to specify."
• Map the existing work. "For this standards area, list which formal bodies, industry consortia, professional associations, and companies already have published work, and tell me where the genuine gaps are rather than where I simply have not looked."
• Classify the standard type. "For each of these practices we want established, tell me whether it should start as a formal standard, an industry guideline, a de facto approach, or an internal company standard, and why."
• Draft the case study. "Turn this implementation experience into a case study suitable as evidence for standards development: the problem, the approach, the challenges, the results, and what did not work."
• Write the internal standard. "Draft internal documentation covering how we evaluate models for fairness and accuracy, document limitations and uncertainty, conduct security testing, manage training data ethically, and monitor deployed models for degradation."
• Test for adoptability. "Review this draft standard against six properties: specific enough to act on, flexible across contexts, evidence-based, inclusive, achievable at varying resource levels, and reviewable. Flag where a smaller organization would be unable to comply."
• Plan the coalition. "Given this proposed standard, identify the types of organization whose support would make it credible, what each of them gains, and what objection each is most likely to raise."

Reflection

Start with the inconsistency question. Think about the practices in your field that every organization does differently with no good reason for the difference: how model limitations get documented, how fairness gets measured, how a vendor claim gets verified. Pick the one that has cost you the most time. That is your standards gap, and you already hold something most participants in standards work do not, which is direct evidence from having lived with the problem rather than having read about it.

Then consider the record question. If you wanted to contribute to standards development in that area tomorrow, what would you bring? Not what you think, but what you could show: a documented internal practice, results from having run it, a case study of what failed first. If the honest answer is that you have opinions and no artifacts, then the first step is not joining a committee. It is writing down what you already do, running it long enough to know whether it works, and publishing the result.

Glossary

• Standard: A formal, often mandatory requirement, developed through consensus and carrying authority once published.
• Best practice: A proven approach that is recommended but not required.
• Guideline: A recommendation for behaviour in a specific context, such as a set of AI ethics principles.
• Formal standards body: An official organization such as ISO, IEEE, or a government agency that develops standards through structured, consensus-based committees.
• Industry guideline: Guidance developed by consortia, professional associations, or non-profits, faster to produce and more flexible than a formal standard.
• De facto standard: A practice that becomes standard through widespread market adoption rather than formal process.
• Company-specific standard: A practice established by a single influential organization that others follow, shaping the market without formal status.
• Standards gap: A critical practice that has not been formalized, evidenced by inconsistent answers across organizations.
• Working group: The committee within a standards body where drafting and revision actually happen.
• Coalition: A group of organizations publicly committing to a standard together, which carries far more weight than individual advocacy.
• Adoptability: The property of a standard being realistically implementable by organizations at varying resource levels.
• Benchmark dataset: A shared dataset used so that model quality can be assessed consistently across vendors and teams.
• Adversarial testing: Testing that deliberately attempts to make a system fail, used as a security and robustness standard.
• Interpretability method: A technique for making a model's decisions understandable, forming part of emerging explainability standards.

Related Lessons

• AI Policy Development for Industry Impact is the internal counterpart: turning standards into policy your own organization runs on.
• Building AI Communities and Industry Networks covers the relationships that coalition-building depends on.
• Creating AI Education Programs for Your Community is how established best practices reach the practitioners who will apply them.
• Regulatory Landscape and Future Compliance addresses the government-led half of the standards landscape.
• Navigating Global AI Regulation goes deeper on jurisdictional variation and what it means for a single standard.
• Bias Auditing and Fairness in AI Systems supplies the measurement practice behind fairness standards.
• Transparency and Explainability in Business AI covers the explainability area in operational detail.
• Ethical AI Frameworks for Business Leaders is the principles layer that guidelines are written from.
• Thought Leadership and Public Speaking on AI is the visibility work that gets a documented practice noticed.

Closing

Standards shape how entire industries operate, and the people who shape them are rarely the people with the most authority. They are the people who noticed a recurring problem, wrote down what they did about it, ran it long enough to have results, and kept showing up where the drafting happened. None of those four moves requires permission, and all of them are available to an organization far smaller than the ones usually credited with setting industry direction.

So begin locally and build outward. Identify the gap from your own experience, research what exists so you are not duplicating it, engage the bodies already doing the work, publish the evidence, and build the coalition that turns a practice into an expectation. Understand that the work is technical and political at once, and favour standards that organizations can realistically implement over ones that are merely rigorous. Over time, the standards you help develop will outlast any individual achievement, multiplying your impact across the industry for decades.

Key Takeaways

• Standards are invisible infrastructure: their value shows up as friction that disappears, which is why their absence is noticed before their presence.
• They deliver five benefits: safety, quality assurance, efficiency, market development, and innovation enablement.
• Standards do not restrict innovation; innovation accelerates once foundations are settled and nobody is rebuilding them.
• Four sources produce standards: formal bodies, industry guidelines, de facto market adoption, and influential company practice.
• Formal standards carry the most authority and move slowest, which means they lag fast-moving fields like AI by construction.
• De facto standards spread fastest and carry the risk of locking in choices that were never optimal.
• Six areas are open in AI right now: responsible AI and ethics, data governance, model evaluation, security and robustness, explainability, and sector-specific requirements.
• The contribution path runs in order: identify the gap, research what exists, engage existing bodies, publish evidence, build coalitions, and evolve best practices into standards.
• Substantive work beats opinion; standards are influenced by those doing the research, implementation, and evidence collection.
• Standards development is political as well as technical, involving competing interests, resource-driven power dynamics, and adoption constraints.
• Effective standards are specific, flexible, evidence-based, inclusive, achievable at varying resource levels, and reviewable.
• Start inside your own organization; documented internal practice is both real influence and the credential for everything beyond it.

Frequently Asked Questions

Why does the AI industry need standards and best practices?

Standards prevent harm, ensure quality, reduce duplicated effort, and build trust. Without them, companies solve identical problems independently, which creates inconsistency and inefficiency. Standards address critical issues including AI ethics and responsible development, data privacy and security, model evaluation and benchmarking, and algorithm transparency. Industries with strong standards attract more investment and talent because risk is easier to price. Standards are not about restricting innovation; they enable it safely and efficiently at scale by establishing baseline practices so that innovation can accelerate above them.

How do industry standards actually get developed?

Standards develop through a combination of formal and informal processes. Formally, official standards bodies such as ISO, IEEE, and government agencies create consensus-based standards through structured committees. Informally, industry groups and consortia develop guidelines through collaboration. Best practices frequently emerge from practice before anyone formalizes them. Effective standards balance rigour with practicality, are evidence-based, go through community review, are accessible and implementable, and evolve as technology changes. The path typically runs from experience to best practices to guidelines to formal standards.

What is the difference between standards, best practices, and guidelines?

Standards are formal, often mandatory requirements, such as ISO 27001 for information security. Best practices are proven approaches that are recommended but not required, such as agile methodology. Guidelines are recommendations for behaviour in specific contexts, such as AI ethics principles. In AI all three matter and they form a sequence. Start with guidelines describing recommended principles, advance to best practices showing how to implement them, and eventually formalize the critical ones into standards. Clear definitions prevent confusion and make the development pathway visible.

How can I influence standards development in my field?

Participate in industry working groups and professional associations, attend standards meetings, and volunteer on committees. Publish case studies and evidence supporting particular approaches, co-author papers and proposals, and work with respected institutions. Build credibility through thought leadership on the topic and document what works in your own organization. Most importantly, contribute substantive work rather than opinions: research, implementation experience, and detailed feedback on drafts. Standards are influenced by those who do the work, not by those who only comment on it.

What are the key AI standards areas emerging now?

The major areas are responsible AI and ethics frameworks covering fairness, transparency, and accountability; data governance and privacy, including GDPR and emerging regulation; model evaluation and benchmarking through standardized datasets and metrics; security and robustness, including adversarial attack testing; explainability and interpretability; and industry-specific standards for sectors such as healthcare and financial services. Standards in these areas emerge through government regulation such as the EU AI Act, international bodies including ISO and IEC, industry consortia, and professional associations. Track those sources to stay current and to identify where to contribute.