Non-verbal prosodic patterns in speech have the power to communicate a speaker’s emotional state, health condition, gender and even personality traits, such as trustworthiness. While research has mainly focused on the relationship between speech acoustics and perceived personality traits from a listener’s perspective, the current research has developed a large speech dataset to examine the production of speech with the intent of sounding trustworthy, based on the speakers’ self-perception. More precisely, the current research is looking to identify whether certain acoustic cues can be used to characterise a speaker’s intent (i.e. neutral or trustworthy). In total, ninety-six younger and older adults from different ethnic backgrounds (i.e. white, black and south Asian) were recruited. They were asked to initially speak a set of sentences in their normal way of speaking (“neutral”) and then repeated the same sentences, but this time they were asked to convey the intent of sounding trustworthy. Our findings provide evidence that pitch and voice quality related features can correctly differentiate a speaker’s intent from our audio dataset with an accuracy of ~70%.