Explainable AI Builds Investor Trust in ESG Forecasts, Indian Study Finds

Artificial intelligence is quietly rewriting the rules of sustainable investing. A new study published in Frontiers in Sustainability shows that when AI systems explain their reasoning and demonstrate strong analytical capability, retail investors are far more likely to trust the ESG forecasts those systems produce — and, ultimately, to prefer sustainable investments. The research, led by M. Vijayananth, R. Muthu Sundari, and N. Saravanabhavan of Vellore Institute of Technology, surveyed 566 middle-income retail investors across four major Indian metropolitan cities and offers one of the clearest behavioral pictures yet of how explainable AI shapes green finance decisions.

The study arrives at a moment when ESG investing has exploded in popularity but remains plagued by skepticism. Investors increasingly evaluate companies not just on financial performance but on environmental stewardship, social responsibility, and governance quality. AI-driven ESG analytics has emerged as a powerful decision-support tool, capable of processing vast datasets to generate forward-looking sustainability insights. Yet many of these systems operate as opaque “black boxes,” leaving investors unable to verify how recommendations are produced. Explainable AI, or XAI, aims to change that by making the logic behind algorithmic outputs transparent and interpretable.

To understand how these technological features translate into actual investment preferences, the researchers built a behavioral framework grounded in Stimulus–Organism–Response theory, supplemented by Signaling Theory and Trust Theory. In their model, two AI-related characteristics — perceived explainable AI and perceived ESG analytics capability — act as external stimuli. These trigger internal cognitive evaluations: first, investors judge whether AI-generated ESG forecasts are credible; then, they decide whether to trust that information enough to rely on it. Only after passing through this sequential mental pipeline do the technological cues influence stated sustainable investment decisions.

The empirical results are striking. Perceived explainable AI exerted a large effect on ESG forecast credibility (f² = 0.559), while perceived ESG analytics capability showed a similarly strong effect (f² = 0.435). Credibility, in turn, significantly boosted trust in AI-generated ESG information, and that trust carried the strongest direct effect in the entire model, driving sustainable investment decisions with a path coefficient of 0.509. Statistical significance held across all hypothesized direct pathways, with p-values below 0.001, confirming that the sequential chain from explainability through credibility to trust is not merely theoretical.

Crucially, the study demonstrates that credibility and trust are not interchangeable. Credibility reflects an investor’s judgment about whether an ESG forecast is believable and reliable; trust reflects the subsequent willingness to act on that information. The researchers found that credibility mediates the relationship between both explainable AI and analytics capability and downstream trust, while trust itself mediates the link between credibility and actual investment preferences. Sequential mediation analyses confirmed that both psychological stages operate in tandem, meaning investors first evaluate whether forecasts appear accurate before deciding whether to depend on them.

But the pathway is not immune to sabotage. Perceived greenwashing risk — investors’ suspicion that companies may exaggerate or misrepresent their sustainability performance — significantly weakened the positive relationship between forecast credibility and trust in AI-generated ESG information. In other words, even when AI systems produce technically credible forecasts, lingering doubts about corporate honesty can erode the trust investors place in those outputs. This finding underscores a persistent vulnerability in sustainable finance: no amount of algorithmic transparency can fully compensate for skepticism about the underlying ESG data feeding the models.

The study also examined whether investors’ own competency in interacting with AI systems moderates the trust-to-decision pathway. Human-AI interaction competency did positively moderate the relationship, with the trust-decision link strengthening from a path coefficient of 0.430 at low competency to 0.587 at high competency. However, the effect size was negligible in practical terms (f² = 0.012), suggesting that while investor skill in interpreting AI outputs matters statistically, it is unlikely to be a dominant driver of sustainable investment behavior on its own. Other contextual and behavioral factors evidently play larger roles.

Methodologically, the research is robust. The team employed purposive sampling with respondent referral controls, collecting data through mall-intercept surveys in Chennai, Mumbai, Delhi, and Kolkata, followed by online questionnaires. Of 978 distributed surveys, 727 were returned and 566 were retained after rigorous screening. All constructs were measured with multi-item scales adapted from established literature, and the model was analyzed using partial least squares structural equation modeling in SmartPLS 4. Reliability, convergent validity, discriminant validity, and collinearity diagnostics all met or exceeded accepted thresholds, and out-of-sample predictive relevance was confirmed through the PLSpredict procedure.

The practical implications are significant for AI developers, financial institutions, and policymakers alike. The authors recommend that AI systems provide understandable explanations for ESG recommendations, clearly identify the factors driving forecasts, and communicate uncertainty rather than presenting outputs as definitive advice. ESG data providers should strengthen validation and document limitations, while financial institutions should incorporate human oversight before presenting AI-generated insights to investors. Policymakers, meanwhile, could reduce greenwashing risk through clearer requirements for AI-assisted ESG disclosure and data verification.

As AI becomes ever more embedded in sustainable finance, this study makes clear that technological sophistication alone is insufficient. What ultimately moves investors toward sustainable choices is a psychological sequence: they must first believe the forecasts are credible, then trust the information enough to act on it. Greenwashing doubts can break that chain, while transparency and analytical capability strengthen it. For an industry betting billions on AI-powered ESG analytics, the message is unambiguous — build systems that explain themselves, and investors will follow.

Subject of Research: Explainable AI and ESG analytics shaping sustainable investment decisions through forecast credibility and trust

Article Title: Explainable AI and ESG analytics for sustainable investment decision: the roles of ESG forecast credibility and trust in AI-generated information

Article References: Vijayananth, M., Sundari, R. M., & Saravanabhavan, N. (2026). Explainable AI and ESG analytics for sustainable investment decision: the roles of ESG forecast credibility and trust in AI-generated information. Frontiers in Sustainability, 7, Article 1965649. https://doi.org/10.3389/frsus.2026.1965649

Image Credits: AI Generated

DOI: 10.3389/frsus.2026.1965649

Keywords: explainable AI, ESG investing, sustainable finance, investor trust, greenwashing, behavioral finance, India, Frontiers in Sustainability, Explainable, analytics, sustainable, investment