A survey - AI in Aerospace: Present and Future

The integration of artificial intelligence (AI) into aerospace engineering has progressed from a primarily experimental discipline to a critical enabler of advanced design, manufacturing and operational capabilities. This paper examines the current applications of AI across the aerospace field and tries to evaluate its predicted evolution in the foreseeable future. At the design level, AI techniques are employed to enhance multidisciplinary optimization processes. Machine learning models are used for simulations and reducing costs. Applications include aerodynamic shape optimization, structural assessment and propulsion system performance. For the moment, these tools are not replacing conventional methods but complement them, especially in early-stage. In manufacturing, AI supports improved process control, defect detection and supply chain management. The adoption of intelligent monitoring systems in composite fabrication and additive manufacturing has contributed to increased reliability and reduced variability in production outcomes. These developments are particularly relevant as aerospace systems incorporate more advanced materials and complex geometries. Operationally, AI plays a central role in autonomy and decision support systems. Current implementations in unmanned aerial vehicles demonstrate the capability of AI navigation, guidance and control algorithms. In commercial aviation, decision-support tools assist flight crews and ground operators in optimizing routes, fuel consumption and maintenance scheduling. In space applications, onboard autonomy supported by AI is essential for deep-space missions, when communication delays limit ground intervention. Predictive maintenance has emerged as one of the most mature applications of AI in aerospace. By analyzing large datasets from onboard sensors and maintenance records, machine learning models can identify degradation patterns and estimate remaining useful life of critical components. This approach improves system availability and reduces lifecycle costs, while maintaining compliance with safety requirements. Despite these advancements, some challenges must be addressed before the adoption of AI in safety-critical aerospace systems can be achieved. Certification frameworks remain largely oriented toward classical systems and are not fully adapted to data-driven models. Issues related to model interpretability, robustness under off-nominal conditions and cybersecurity must also be resolved. Furthermore, the integration of AI into legacy systems requires careful consideration of system architecture and human-machine interaction. Looking ahead, the continued development of AI in aerospace is expected to focus on hybrid modeling techniques, increased onboard autonomy and integration between digital design, manufacturing and operational feedback. Progress in these areas will depend on advances in computational resources, data availability and regulatory adaptation. This paper provides a structured overview of these topics, linking current capabilities with foreseeable developments and identifies key research directions necessary to ensure the safe and effective integration of AI into future aerospace systems.