Artificial intelligence is reshaping every stage of mechanical engineering, from modeling, simulation and optimization to diagnostics, design and technical communication. This paper draws on a round‑table at the 28th ABCM International Congress on Mechanical Engineering (COBEM 2025) to offer a structured view for Brazilian and global ME communities. Rather than a literature survey, it focuses on how AI can improve practice while preserving scientific rigor.
We first separate statistical learning from human cognition, describing current AI capabilities, limitations and the emerging agentic forms. Then we examine pathways for industrial adoption, workforce shifts, AI‑assisted discovery, and the energy and infrastructure demands of AI computation, highlighting implications for sustainability and thermal management.
In safety‑critical applications, physical knowledge, structured reasoning and domain expertise must be embedded in AI systems. Relevant techniques include physics‑informed hybrid learning, physically constrained modeling, validation, uncertainty quantification and certification. Educational challenges involve curriculum reform, human‑agent collaboration, professional accountability and the preservation of engineering knowledge.
The article concludes with principles for responsible AI use: physical consistency, rigorous validation, explicit uncertainty treatment, energy sustainability, human oversight and ethical responsibility. AI should complement, not replace, engineering judgment.
Review