AI Fraud Report 2026, Only 7% of Companies Ready

From a sectoral perspective, the research analyzed the responses of specialists from over twelve production sectors, finding a greater concentration in the public sector, which represents 26% of the sample, and in the banking and financial services sector, at 23%. The data indicates that every type of crime supported by artificial intelligence has recorded an increase in the last two years.

The most relevant phenomenon is applied social engineering to deepfakes, reported as growing by 77% of those interviewed, followed by direct consumer scams and document falsification using generative artificial intelligence, both estimated to be increasing by 75% of the sample. Projections for the next twenty-four months confirm this trend, with 55% of professionals predicting a further increase in document falsification and video and anthropomorphic manipulation.

The discrepancy between the development of threats and response capabilities raises structural and regulatory questions for businesses. John Gill, President of ACFE, analyzed the current scenario: “The data paints a worrying picture: Fraud is evolving faster than most organizations can defend themselves against. AI-based threats are not a future risk, but they are already a reality and are growing quickly. Industry specialists have made real progress in the adoption of AI, and what has emerged in this report is a wake-up call. Companies that fail to strengthen their defenses against AI-enhanced fraud risk becoming even more attractive targets”.

The adoption of machine learning and artificial intelligence tools in corporate law enforcement programs shows a steady progression, rising from 18% in 2024 to 25% in 2026, with an additional 28% of organizations planning to implement such protocols by 2028. However, the analysis highlights a delay in the governance and validation processes of the models. Although 86% of the sample considers the accuracy of the results to be a crucial factor for the integration of generative artificial intelligence, only 18% carry out systematic tests to identify statistical biases or fairness problems in the models. Furthermore, compared to 82% of respondents who consider the explainability of algorithms an absolute priority, only 6% declare they possess the necessary tools to motivate the decisions made by the models. This deficiency translates into potential regulatory fines and legal liabilities, particularly for regulated industries such as banking and insurance.

On the economic level, 55% of companies expect an increase in budgets allocated to protection technologies in the next two years, although financial constraints remain the main structural obstacle to implementation for 84% of those interviewed. The temporal dynamics linked to the approval of investments directly influence the risk exposure of companies. Stu Bradley, Senior Vice President of Risk, Fraud and Compliance Solutions at SAS, highlighted the operational asymmetry between companies and criminal organizations: “Cybercriminals have no governance committees and don’t wait for budget cycles or regulatory clarity – they just act. Every quarter that business leaders spend evaluating a technology is three more months in which lawbreakers can weaponize it and find organizations unprepared”.

The evolution of emerging technologies shows a gradual transition of generative artificial intelligence from an exploratory phase to a practical application. Currently used by 16% of professionals, the technology sees future interest from 58% of interviewees, mainly focused on the detection of phishing activities, quantitative risk assessment and internal reporting. Systems based on autonomous agents have the highest expected growth rate, with a current usage share of 8% and adoption forecasts of 31% by 2028.

Physical biometrics consolidates its lead as the most widespread emerging technology, reaching 45% of preferences compared to 34% in 2022in the face of still limited use of native cloud platforms and process automation systems. Finally, 62% of the sample estimates that the Quantum computing and quantum artificial intelligence will have a major impact on crime prevention by 2030with 11% of experts detecting the first operational effects already in the current context. The complete analytical data and the related enforcement methodologies remain accessible within the programmatic document which can be consulted at SAS.com/fraudreport.