8The modern tech industry has been shaped by several waves of technology. The first was the shift from memory to processing in the 1980s. The second was the mainstreaming of the internet, which fed the dot-com era. The third was the adoption of social media, which turned online identities into multi-layer networks. The fourth was the migration to the cloud for efficiency and redundancy. Together, these waves rapidly created new customers for online services and, more importantly, enabled the free movement of ideas, goods, and technologies. This dynamic produced the success stories of Google, Facebook, Amazon, and LinkedIn, all of which rely on connectivity, engagement, and growth. 

The current AI wave is the next cycle of this phenomenon, promising to change how we work and live, powered by enormous volumes of data and massive processing capability. 

The bad news is that the mistakes of the past, including data breaches, inappropriate access,  disinformation, and non-consented use cases, will carry much bigger consequences this time around, for consumer trust, for regulatory compliance, and for the AI-powered services themselves. 

“Responsible AI” is not a “do the right thing” altruistic imperative. It is a critical vector that will  determine which frontier models gain market share, which companies produce sustainable ROI,  and whether AI actually delivers the financial returns that investors seem to have already priced  

into AI company valuations. At its core, responsible AI is the practice of earning trust through data handling, model behavior, AI governance, and transparency. 

Just as the era of cheap venture capital for internet and dot-com companies did not last forever,  neither will today’s era of light-touch, low/no AI regulation. Delays in data hygiene compound over time, and AI will produce negative outcomes along the way. The result will be regulators and courts looking for ways to hold AI companies accountable. 

The way forward is building governance models for AI that will deliver more accountable AI  services AND produce better business outcomes. 

The numbers paint a striking picture. 

“Only 6% of companies fully trust AI agents to autonomously run their core business  processes.”

According to Informatica, these are the barriers for businesses deploying agentic AI (with the  percentage of companies citing each barrier): “data quality/retrieval concerns (50%), security concerns (43%), observability concerns (39%), and lack of safety guardrails for AI use (35%).” 

The consumer numbers are just as stark. A Pew Research Center survey of over 5,000 U.S. adults, published in June 2026, found that 71% of Americans believe increased AI use will make their personal information less secure, and 67% have little to no confidence in government’s ability to regulate AI effectively (up from 62% just two years earlier).

A December 2025 YouGov poll found that only 5% of Americans say they trust AI “a lot,” while 41% say they actively distrust it. A Quinnipiac University poll from March 2026 found a similar pattern even among people who use AI regularly: 51% of U.S. adults now use AI for research, but only 21% trust the information it gives them most or almost all of the time. 

The 2026 Edelman Trust Barometer Special Report found that 88% of consumers globally say it’s a critical deal-breaker. Trust is on par with the two other leading purchase factors: a brand offering the best quality (89%) and good value for money (88%). For AI  products specifically, the impact of trust may be amplified even more because an AI system that outperforms on cost and capability could still lose the sale if the buyer fears it will leak data,  hallucinate in production, or expose them to other regulatory risk. 

Responsible AI practices aren’t just a defense against reputational or legal damage; they’re now a prerequisite for closing the deal in the first place. 

Source: 2026 Edelman Trust Barometer Special Report: Brand Growth in an Insular World, p. 11. 

Two tech companies that have earned trust, and consequently built market share, are Apple and  Anthropic. 

Perhaps no company better illustrates the business case for building trust on ethical principles such as data privacy than Apple. While Google and Facebook were under fire for exploiting user data, Apple CEO Tim Cook made privacy a central brand promise with a "Privacy Matters" campaign spanning prime-time ads and over 10,000 billboards worldwide. 

In 2021, Apple’s App Tracking Transparency feature, an opt-in dialog giving users control over cross-app tracking, cost Meta about $10 billion in annual advertising revenue, while Apple’s stock rose more than 37% that year. Privacy wasn’t a constraint on Apple’s growth; it was an accelerant. Investments in on-device AI processing both strengthened user privacy and justified premium hardware pricing, creating a self-reinforcing cycle that competitors relying on cloud-based data collection cannot easily replicate.

Data backs this strategy: 84% of consumers say they’re loyal to companies that protect their privacy, and 97% of companies report gaining competitive advantage or investor confidence from privacy investment.

Yes, Apple collects an estimated $20 billion a year from Google for making Google the default search engine in Safari, Siri, and other services. This business avoids Apple monetizing user data directly, but could reasonably be accused of monetizing user attention instead, just one layer removed. Nonetheless, Apple’s primary revenue comes from expensive hardware its customers choose to buy, not from targeted ads that use data extracted from customer behavior. That’s a meaningfully different position from companies whose entire business model depends on harvesting and reselling user data, even if it isn’t the fully data-agnostic story that Apple’s marketing campaigns imply. 

Amidst the AI trust gap, Anthropic has earned trust by embracing safety as a core principle. This headline stands out: “Anthropic’s Safety First Approach Won Over Big Business.”  

Anthropic’s founders left OpenAI over disagreements about safety priorities. Putting responsible  AI development ahead of speed to market became the company’s defining competitive differentiator. Its approach of “Constitutional AI” embeds a written set of ethical principles directly into the model’s training process. 

Market data is consistent with what the trust gap would predict, even if it can’t prove the direct causal link on its own. According to Menlo Ventures’ 2025 State of Generative AI in the  Enterprise report, Anthropic commanded 40% of enterprise LLM API market share by the end of  2025, more than triple its 12% share in 2023. Google climbed to 21%, a threefold increase from  7%. Meanwhile, OpenAI’s share fell from 50% to 18%. 

Anthropic’s revenue grew from roughly $1 billion to over $5 billion in annualized run-rate in the first eight months of 2025 alone. The company counts eight of the Fortune 10 among its customers. In December 2025, Anthropic announced a partnership with Accenture to train  30,000 professionals on Claude, explicitly grounded in a “shared commitment to responsible  AI.” For regulated industries like healthcare, finance, and law, where compliance is non-negotiable, Constitutional AI is one reason enterprises choose Anthropic over the alternatives. 

As Cognizant’s Chief Responsible AI Officer put it, “By operationalizing responsible AI and demonstrating it with evidence, organizations can scale faster, meet cross-border requirements,  and convert trust into competitive advantage.” 

The most public test of those principles came in early 2026. Anthropic had signed a $200 million contract with the Pentagon in 2025, with the Department of Defense accepting Anthropic’s usage restrictions as a condition of the deal. Negotiations later stalled when the DoD sought unrestricted use of Claude “for all lawful purposes,” including mass domestic surveillance and fully autonomous weapons systems operating without human oversight. Anthropic held its position. The Pentagon responded by designating Anthropic a “supply chain risk” and directing military contractors to stop using its products. Anthropic sued in federal court and, in March 2026, obtained a preliminary injunction. Most recently, on August 27, 2026, a U.S. District Judge ruled that the DoD’s blacklisting of Anthropic was unlawful.

The Pentagon dispute may have cost Anthropic government revenue in the short term. But public response and business data tell a different story. Claude downloads surged, and within days it was the #1 free app downloaded in the Apple App Store, dethroning ChatGPT. The number of daily users climbed by 70%, and Anthropic’s brand positivity score increased by 14.6 percentage points. 

That same trust calculus is now reshaping how enterprises actually buy AI, including adjusting their vendor selection criteria. Enterprises are more carefully evaluating “reliability” and  “security,” which are both words for trust in operational form. Our research found that enterprise  buyers now use “reliability” as a catch-all for three things: 

Enterprises are also not investing in secure AI just to “do the right thing.” IBM’s Cost of a Data Breach 2025 Report found that the average breach now costs $4.44 million, and 97% of organizations reporting AI-related incidents lacked adequate access controls. “Shadow AI,” with employees using unauthorized AI tools, adds another $670,000 to the average breach cost. In an environment where reputational risk is existential, responsible governance functions as the surest form of insurance. 

Despite awareness of responsible AI’s measurable benefits, the World Economic Forum found that 81% of companies remain in the early stages of implementation. While this governance gap puts current AI investments at risk, it also means companies that act now should gain a distinct competitive advantage. 

Whether you’re building AI systems or integrating them into your operations, you’re more likely to succeed when trust, transparency, and security sit at the core of your AI strategy. In an uncertain world, deploying responsible AI isn’t just the right thing to do. Evidence shows it’s the best business strategy.

The AI Collective is built by volunteers across 180+ chapters in 40 countries.

Thank you to the thousands of volunteers around the world who make this work possible. We truly could not do this without you.