Novel foods have been the object of intense public debate in recent years. Despite efforts to communicate the outcomes of risk assessments to consumers, public confidence in the management of potential risks associated has been low. Various reasons behind this has identified, chiefly a disagreement between technical experts and consumers e.g. over the nature of the hazards on which risk assessments should focus and perceptions of insufficient openness about uncertainties in risk assessment. The consumers part of the EU-project, NOFORISK, investigate the disagreement by comparing laypeople and experts understanding of benefits and risks associated with three Novel foods (a potato, rice and functional food ingredients) using a relatively new methodology for the study of risk perception called Mental models. Mental models focus on the way people conceptualise hazardous processes and allows researchers to pit a normative analysis (expert mental models) against a descriptive analysis (consumer mental models). Expert models were elicited by means of a three-wave Delphi procedure from altogether 24 international experts and consumers models from in-dept interviews with Danish consumers. The results revealed that consumers´ and experts' mental models differed in connection to scope. Experts focused on the types of hazards for which risk assessments can be conducted under current legal frameworks whereas consumers were concerned about issues that lay outside the scope of current legislation. Experts' defined risk and benefit in terms of detailed chains of cause-effect relationships, but consumers used abstract concepts when they reasoned about biological processes. Outcome uncertainty played an enormous role in consumers' perception of risk, which was in contrast to experts, who often declined to elaborate on consequences that in nature were undefined. Such states of ignorance appeared to be the main drivers behind consumers' concerns. A subjectively safe history of personal use turned out to be the main barrier as consumers actively defended their personal eating history, maintaining the security offered by lifelong habits. Consumers found it utterly unconvincing that, all of a sudden, they should regard their everyday foods as toxic and therefore it might not be possible to effectively communicate the health benefits of some novel foods to consumers. Several misconceptions became apparent in the consumer data as consumers 1) implicitly assumed that the novel foods had substantially improved agronomic properties and 2) assumed that all novel foods were governed by the same body of legislation that applies to GM foods. The last misconception might influence consumer trust in risk management when consumers realize that novel foods other than GM foods do not have to undergo environmental risk assessment. Another implication for risk management appeared as consumers did not demand any participatory elements in the risk analysis process. Consumers talked extensively about normative, governance-related issues during the in-depth interviews but did not mention stakeholder involvement or public consultation mechanisms. Though participatory elements may be desirable in their own right from a normative, political point of view, they might not be more effective when used as instrument for increasing consumer trust in the legitimacy of the risk analysis process.