Brand AI faces trust gap: 56% find it misleading; only 29% say it feels human

New Delhi: Brands may be embedding AI across advertising, commerce and customer service, but consumers are not necessarily convinced by the experiences being created. A global study found that 56% of consumers believe the use of AI by brands can feel misleading, while only 29% say it feels human. Half of respondents also said AI can make brands look, feel or sound the same. The findings come from research by Havas across seven markets, including India, examining how consumers respond to AI when it appears in advertising, product discovery, customer support, recommendations and other brand interactions. The gap is emerging even as AI becomes an increasingly normal part of consumers’ lives. Across the seven markets, 76% of respondents said they use AI sometimes or regularly. About 39% use AI tools to research products, services or brands before making a purchase decision, while 42% said AI recommendations can help them discover new products, services or brands. But higher usage has not translated into unconditional trust. Only 35% said they trust brands to use AI responsibly, with the number falling to 18% in the US. Separately, 65% said they were concerned that brands using AI could expose them to false or misleading information. Half think brands may be cutting corners The research suggests that consumers respond differently to AI they choose to use themselves and AI deployed by brands. Around 41% said they find AI exciting when they choose to use it, compared with 32% when a brand uses AI to interact with them. Half of respondents said using AI can make brands feel as though they are cutting corners. Another 50% said brand use of AI can feel cold. Customer service also remains a source of friction. About 67% agreed that AI-powered customer service can be frustrating when users feel stuck in a loop, while 57% said they use AI to find information but still double-check important details. Questions about commercial intent are also influencing trust. More than half, or 56%, said that when an AI chat tool recommends a brand, they assume the company may have paid for the recommendation. Consumers want brands to disclose AI use Transparency emerged as one of the clearest expectations from consumers. Around 78% said it is important for brands to clearly disclose when AI is being used. The study found that consumers were more receptive when AI-generated material was labelled and when outputs appeared grounded in reality rather than overly polished or artificial. Data control is another concern. About 74% said people should be allowed to opt out of AI-driven personalisation, while 53% said they would share data for personalisation only if they could control how it was used. Consumers also expect brands to take responsibility for the systems they deploy. Nearly three-fourths, or 73%, said brands should actively test AI systems to ensure they do not reinforce bias, while 69% said systems should be checked for reinforcement of stereotypes. India more receptive to brand AI India stood out as one of the more receptive markets in the study. On a 10-point scale, India recorded the highest average score for Utility at 6.89, ahead of Brazil at 6.53 and China at 6.10. Utility measures whether an AI-enabled experience feels useful and relevant. India also recorded the highest Credibility score at 6.62, followed by Brazil at 6.16 and China at 5.81. The US scored lowest at 4.11. Credibility measures whether an experience feels believable and trustworthy. On Appeal, which measures whether an AI experience feels attractive and exciting, India again ranked highest at 6.83. Brazil followed at 6.13 and China at 5.62, while the US stood at 3.23. The India scores suggest greater openness to AI-enabled brand experiences than in several Western markets, although the wider global findings show that usefulness does not automatically translate into trust. AI performs better when it has a clear job The study also found that consumers respond better when AI helps them accomplish a task rather than when it is primarily being used to sell or persuade. AI-assisted education and support received the highest overall score at 56.7 out of 100, followed by AI-assisted product comparison at 55.4 and brands appearing in AI chat tools at 54.9. At the other end, influencer or creator-style AI content scored 45.9, followed by brand advertising at 48.4 and dynamic or shoppable advertising at 49.6. Automotive was the highest-scoring sector overall at 55.3, followed by insurance at 53.7 and fashion at 53.3. Grocery scored 50, while healthcare OTC was lowest at 48.2. The research covered 10,558 respondents across India, Brazil, China, France, Germany, the UK and the US, and evaluated more than 21,000 AI-enabled brand experiences across five sectors and seven consumer touchpoints. Havas developed its own scoring framework from the research to assess AI experiences on utility, credibility and appeal while accounting for perceptions that an experience may be misleading or manipulative. Only around 9% of the AI experiences evaluated reached the highest scoring band, indicating that widespread use of AI has not yet translated into consistently strong consumer experiences.