AI, in combination with the analytics, It has been evolving rapidly towards playing a relevant role in different stages of the evaluation processes, allowing the analysis of large volumes of information, the identification of patterns and the making of predictions not only about future performance but also about risks (such as turnover), the identification of performance gaps and the planning of preventive interventions, before the consolidation of problems. Additionally, it allows going beyond the point-in-time assessment of competencies, enabling an integrated and longitudinal analysis of potential and fit of talent. Through the collection and processing of the data obtained in assessment, Predictive models can be constructed that estimate the probability of future success in certain organizational functions or contexts.

These approaches contribute to reducing the subjectivity traditionally associated with potential assessment, allowing for consistent comparison of talent, identifying relationships between potential assessment and actual performance, and supporting strategic decisions regarding selection, succession, development, or mobility. Furthermore, dashboards and automated reports facilitate the visualization and communication of results, promoting greater transparency associated with decisions and the monitoring of metrics.

However, the effectiveness of analytics It depends heavily on the quality of the data, the proper definition of indicators, and the critical interpretation of the results. Without these precautions, there is a risk that decisions based on irrelevant or biased metrics will be made.

Regarding AI, currently its use in assessment occurs mainly in three dimensions (HR Daily Advisor, 2025). The first refers to the assessment of technical skills through tests that integrate AI and automated interviews, where AI emerges as a partner that supports the recording and summary of the responses shared by the evaluated, ensuring standardization and efficiency. The second corresponds to the assessment of cognitive skills, such as logical reasoning, numerical, verbal, and inductive reasoning, with automatic correction and reports that claim to have high statistical rigor. Finally, AI is also beginning to be applied to the assessment of behavioral skills, using natural language processing, digital simulations, and behavioral data analysis to infer skills such as communication, adaptability, or decision-making. However, the assessment always depends on human supervision, since all decisions concerning people must be made by humans. Indeed, this use of AI – in recruitment and talent assessment contexts – is the subject of growing regulatory attention in multiple jurisdictions. This move reflects the need to balance technological innovation with the protection of the fundamental rights of workers and candidates. First and foremost, the European Union is at the forefront of AI regulation globally with Regulation (EU) 2024/1689, also known as EU AI Act, The new regulation, which came into effect in August 2024, adopts a risk-based approach, classifying AI systems into four levels: unacceptable risk, high risk, limited risk, and minimal risk.

Now, despite the benefits, significant risks have been identified. One of the main ones is the “black box” effect, in which the complexity of algorithms makes it difficult to understand the underlying criteria for potential decisions. insights automated. Moreover, the systems can reproduce biases present in the historical data used in AI training, perpetuating associated discriminations, for example, based on gender, ethnicity, or age.

The current state of assessments It is characterized by an increasing tension between technological efficiency and the authenticity of evaluation. AI has become an unavoidable reality for organizations and individuals, making its complete exclusion from the processes unrealistic. Thus, we look at the possibility of using hybrid models, that is, models that combine automated evaluations where AI can be used to facilitate the work of human resources professionals, as well as data analysis without neglecting human interpretation and validation.

In parallel, ethical and legal issues are becoming increasingly important, namely the protection of privacy, the control of algorithmic bias, and the guarantee of equal opportunities for those evaluated. It should be noted that the EU AI Act specifically prohibits certain practices considered unacceptable risks, namely: the recognition of emotions in work and educational contexts, social scoring systems (social scoring), biometric categorization to infer sensitive characteristics such as race, political orientation or religious beliefs, and subliminal manipulation techniques.

The future of assessment The potential of AI will be marked by the consolidation of the IA–Analytics binomial as the central axis of decision-making in Human Resources. The challenge will be both technological and conceptual: defining which data to collect, how to interpret it ethically, and how to balance efficiency with justice and human conscience. The regulatory trend points to a growing international convergence around principles of transparency, non-discrimination, and human oversight in the use of AI in work contexts. Organizations operating in multiple jurisdictions must anticipate the application of the most demanding standards, assuming that the requirements of EU AI Act they will gradually become the benchmark global.

Laura Andries, Consultant, SHL Portugal
Carolina Cordeiro dos Santos, Senior Associate, Vieira de Almeida

Published in Human Magazine on 03/03/2026