Artificial intelligence (AI) is reshaping scientific research by accelerating discovery, improving efficiency, and enabling new uses of complex data, while raising ethical chal lenges involving accountability, explainability, privacy, data integrity, and bias. In research, these concerns matter because scientific validity, participant protection, and public trust are directly at stake. This manuscript argues for an ethics-by-design approach in which ethical reflection, challenge, and decision-making are embedded from the earliest stages of AI-enabled research, rather than treated as a downstream compliance exercise. Aligned with international guidance on responsible AI, the article proposes a practical framework organised around four reinforcing dimensions: people, process, governance, and communities. Because AI systems are dynamic and context-dependent, this model calls for adaptive oversight, cross-functional accountability, and meaningful stakeholder engagement. By integrating ethics into design, deployment, and review, research organisations can better balance innovation with responsibility, helping ensure trust worthy, socially legitimate scientific progress.