Trust is no longer a defensive strategy; it's an offensive advantage.
AI is everywhere, but anti-AI sentiment is rising. Why? Because trust hasn't kept pace with adoption. Now in its second year, The Data and AI Impact Report tracks how organizations around the world are adopting AI, building trust in AI and realizing business value from it.
This year’s research shows that reality is setting in. As organizations move from generative AI to more autonomous systems, trust is becoming the factor that separates success from stagnation.
Focus on explainability
One year ago, trust was primarily a question of generative AI. Today, organizations are evaluating whether they trust AI to act on their behalf. As AI moves from generating output to making decisions and taking action, trustworthiness becomes increasingly important. Agentic AI now sits at the center of the conversation, and trust is struggling to keep up.
GenAI is trusted by 76% of respondents. Agentic AI, however, is only trusted by 66%. A 10% point drop. Why? Because trust drops as autonomy scales.
The drop in trust has consequences. 97.2% of respondents say they override at least some AI recommendations. The leading reason isn't poor output quality. It's a lack of explanation and context. When people can't understand how a decision was reached, they're less willing to let AI act on their behalf.
Trust Declines as AI Autonomy Increases
Trust Generative AI
Trust Agentic AI
Build better data
AI continues to have a data problem. Year two of this research shows that most organizations have data infrastructures that can’t fully support the explainability, lineage of validation that agentic AI requires.
Only 17.5% have fully optimized data infrastructure (up from 10.2% last year) that provides the transparency needed to trust agentic decisions.
When you can show the lineage of a decision, you can defend it. When you can’t, you land in the large group who are overriding decisions and, in turn, wasting their investment.
The full report breaks down how data maturity is increasing, but not fast enough for AI.
Data Maturity Remains a Major Barrier
Optimized Infrastructure Today
Optimized Infrastructure Last Year
Follow the leaders
What is a trustworthy AI leader?
This research measures the trustworthiness of AI across five dimensions: Data Quality & Governance; Model Governance & Oversight; Explainability & Fairness; Responsible AI Policy; Audit & Accountability.
Together, these dimensions form the Trustworthiness Index. Organizations scoring 80 or higher are classified as trustworthy AI leaders.
The biggest lesson from two years of research: trustworthiness and business outcomes move together. The organizations investing most aggressively in trustworthy AI continue to separate themselves from the rest of the market. Trustworthy AI leaders are 15x more likely to report strong ROI, but the difference goes beyond financial performance. Leaders are more likely to invest in governance, explainability, validation and employee training, creating systems people are willing to trust and use.
Those leading in responsible AI, and by extension ROI, are making moves to extend that lead: 85% are growing their trustworthy AI investment by 10% or more next year vs only 26% of organizations falling behind. The gap between leaders and laggards is already significant. As leaders continue increasing investment in trustworthy AI, that gap may become more difficult to close.
Planned Growth in Trustworthy AI Investment
High-Trust Organizations
Low-Trust Organizations
Industry snapshot
While every industry faces different regulations, risks and priorities, the same pattern appears throughout the research: organizations that invest in trustworthy AI consistently outperform those that don't.
Choose your industry and explore more details
Banking
Highlight
Banks are extending decades of risk management and compliance practices into AI, helping build trust and drive stronger results. 85% of trustworthy AI leaders in banking have rolled out responsible AI policies to every employee, compared to just 29% of laggards.
Key concern
As AI becomes more autonomous, oversight remains a challenge. 38% of trustworthy AI leaders cite managing autonomy and human oversight as a concern, compared with 12% of laggards.
Responsible AI Policy Adoption Drives Trust in Banking
High-Trust AI Organizations
Low-Trust AI Organizations
And high-trust banks are 6.6x more likely to report strong AI ROI.
Insurance
Highlight
Insurance provides one of the clearest examples that trustworthy AI and business performance go hand in hand. 59% of trustworthy AI leaders report strong or high AI ROI. Among laggards, that number drops to 0%.
Key concern
As AI scales, maintaining the right balance between autonomy and oversight becomes increasingly important. 34% of trustworthy AI leaders identify managing autonomy and human oversight as a key concern.
High AI Trust Drives Stronger ROI in Insurance
High-Trust Insurers
Low-Trust Insurers
Life Sciences
Highlight
Life sciences organizations are building on established quality, audit and compliance practices to strengthen AI governance. 91% of trustworthy AI leaders have responsible AI policies that reach every employee, compared with only 7% of laggards.
Key concern
The data foundation remains critical. 62% of trustworthy AI leaders cite data privacy as a concern, reflecting the sensitivity and complexity of the data they manage.
Responsible AI Policies Reach Every Employee
Higher-Trust Organizations
Lower-Trust Organizations
Public Sector
Public sector organizations are increasing investment in trustworthy AI as they expand AI use across citizen-facing services and decision support. 41% of trustworthy AI leaders plan to increase trustworthy AI investment by more than 20% next year, the highest rate of any focus industry.
Priority focus
Transparency remains a key hurdle. 41% of trustworthy AI leaders identify explainability and transparency as a remaining gap, compared with 26% of laggards.
41% of AI Leaders Plan Significant Increases in Trustworthy AI Spending
Increase in Trustworthy AI Investment
The highest rate of any focus industry.
With trust gaps narrowing across industries like life sciences and financial services, the research highlights the practices that drive real impact in regulated environments. It’s an essential perspective for industries where oversight is paramount, and strengthening trust and governance is key as AI becomes more autonomous. Mona Chadha Director of AI & Strategic Partnership Amazon Web Services
More on Trust in AI
Connect this year's findings to the broader movements transforming business and technology.
