AI adoption is surging while trust falls behind

More than 50% of users want control over their AI data

AI adoption is rising faster than trust in how providers handle personal data. More than half of respondents to Tether’s survey said they want to own the data created through their AI interactions. Yet only 35% were confident that their personal data was secure.

That gap matters. People now use cloud-based tools such as ChatGPT and Claude for research, problem-solving, and personal advice. These interactions can involve sensitive information. Because cloud AI typically sends user inputs to remote infrastructure for processing, data governance becomes part of the product experience.

The main constraint is trust. Users can value an AI service while remaining uncertain about where their information goes, how long it is retained, or how much control they have over it. Higher adoption therefore does not prove that users are comfortable with current data practices. The Tether findings point in the opposite direction: frequent use can coexist with low confidence in data security.

For executives, data ownership and protection should be treated as product requirements. Clear retention policies, meaningful user controls, secure processing, and simple explanations of how data is used can reduce uncertainty. These measures also matter commercially. AI products increasingly depend on users sharing useful context, and users who do not trust the system may limit what they provide.

The strategic priority is clear: companies need to make AI useful without requiring customers to accept unclear data practices. As AI becomes embedded in more products and workflows, trust will increasingly affect adoption, customer relationships, and the types of information users are willing to share.

More than 50% of users will trade some privacy for AI benefits

Privacy concerns are not stopping people from using AI. Tether’s survey found that more than half of respondents would give up some privacy for AI-enabled convenience or share personal data in exchange for lower-priced AI services.

This finding exposes an important distinction between consumer attitudes and behavior. Users can be concerned about data security while still deciding that an AI service provides enough value to justify sharing information. Convenience and price can outweigh privacy concerns at the point of use.

For executives, this does not mean privacy matters less. It means customers assess privacy alongside tangible benefits. A service that saves time, removes repetitive work, or costs less can encourage users to accept more data collection. But that willingness has limits, especially when the information involves health, financial, identity, or other sensitive data.

The business risk is treating user consent as proof of trust. A customer may agree to data collection because access to a useful feature depends on it. That decision does not necessarily mean the customer understands or supports every subsequent use of the data. Companies should therefore make the exchange explicit: what information is collected, why it is needed, how it is processed, how long it is retained, and what controls the user has.

The stronger strategy is to compete on both value and data discipline. AI services can remain convenient and affordable while collecting only the information required for a defined purpose. Firms that give customers meaningful choices can preserve the benefits driving adoption without making broad data access a default condition of participation.

AI attitudes differ sharply by market

AI adoption does not translate into the same expectations everywhere. Tether’s survey found greater skepticism among users in the U.S. and UK, including stronger concerns about the privacy of health and medical data. Respondents in India and Nigeria were more optimistic about AI and showed greater interest in functionality that works across devices.

For global companies, these differences matter because a single AI product and privacy strategy may not satisfy every market. A user willing to connect an AI service across a phone, computer, and other devices may value seamless access. Another user may place greater value on limiting where sensitive information is stored and processed. Product design needs to account for both priorities.

The constraint is not simply consumer sentiment. Privacy rules, expectations around consent, and sensitivity to particular types of data also vary across jurisdictions. Health information deserves particular attention because misuse or exposure can carry significant consequences. Executives should therefore distinguish between features that can remain global and controls that need to reflect local requirements.

The survey results should not be used to label entire countries as either pro- or anti-AI. They show relative differences among surveyed respondents, not uniform attitudes within national populations. Factors such as age, income, digital access, professional use, and familiarity with AI can also influence how individuals assess its benefits and risks.

For multinational businesses, localization should extend beyond language and marketing. Companies should determine which data must move between devices, which processing can remain local, what users can control, and how consent is presented in each market. Regional differences in trust can then inform product architecture and governance rather than becoming a barrier to expansion.

On-device AI could turn privacy into a product advantage

Rising AI use and low confidence in data security create a clear opening for on-device AI. Instead of sending every request to remote cloud infrastructure, local AI can process some or all information directly on a user’s phone, computer, or other device. Sensitive data can therefore remain under tighter user control.

The demand for this model is growing as users become less comfortable with cloud-based AI. This connects directly with Tether’s survey findings: more than half of respondents want ownership of data created through their AI interactions, while only 35% are confident their personal data is secure. Local processing can address part of that trust gap by reducing the amount of information that needs to leave the device.

The technical constraint is performance. On-device systems operate with less computing power and memory than large cloud data centers. They can also face limits in model size, battery consumption, updates, and access to current information. A local AI product therefore needs to deliver sufficient speed and capability without creating a poor user experience. For many businesses, the practical answer may be a hybrid design in which sensitive or routine tasks run locally while workloads requiring greater computing capacity use the cloud.

This approach also changes the economics and governance of AI. Local processing can reduce dependence on continuous cloud inference and limit the amount of customer information transmitted to external infrastructure. But it does not remove the need for security. Companies still need to protect models, stored information, software updates, device access, and any data that moves between local and cloud systems.

Paolo Ardoino, CEO of Tether, presents private local AI as a route to broader adoption. He said: “By building options for local, private AI tools that can match the speed and power of current market offerings, we can raise the ubiquity of AI technology while doing away with barriers that discourage new users.” He added: “The future of society can thrive only if AI becomes open and accessible to everyone.”

For executives, the opportunity is larger than adding a privacy feature. Data location, processing architecture, and user control can become core product decisions. Companies that can move suitable AI workloads onto devices while retaining competitive performance could reduce privacy concerns and give customers greater control without sacrificing the utility that drove AI adoption in the first place.

Key executive takeaways

• Trust is lagging AI adoption: More than half of surveyed users want ownership of their AI data, while only 35% are confident their personal data is secure. Leaders should treat clear data controls and secure processing as product requirements.
• Users will trade privacy for value: More than half of respondents would accept some privacy loss for greater AI convenience or lower prices. Companies should deliver these benefits while minimizing unnecessary data collection and making consent clear.
• AI expectations vary by market: U.S. and UK respondents showed greater skepticism and health-data privacy concerns, while users in India and Nigeria were more optimistic and interested in cross-device features. Global AI strategies should adapt privacy controls and product design to regional expectations.
• On-device AI creates a privacy opportunity: Processing more AI workloads locally can reduce cloud data exposure and give users greater control. Leaders should assess local and hybrid AI architectures based on privacy, performance, cost, and workload requirements.