This paper examines the ethical status of face-to-face communication in clinical practice, using autism as an illustrative case. Drawing on phenomenological accounts of embodiment and first-person descriptions of autistic experience, it argues that face-to-face interaction-characterized by immediacy, multimodality, and unpredictability-can impose significant perceptual and cognitive burdens on autistic patients. Digitally mediated communication, by contrast, provides greater temporal flexibility, semiotic stability, and user control, thereby reducing communicative pressure and supporting more accessible participation. Reframing communication through the principles of autonomy, non-maleficence, and relational care, the paper argues that communicative practices should be understood as context-sensitive and responsive to patients' perceptual conditions rather than governed by a default preference for face-to-face interaction. It concludes that presence should be reconceptualized as a relational achievement grounded in responsiveness rather than physical co-presence alone, with implications for how ethical communication in clinical care is understood.
Organoid-artificial intelligence (AI) platforms are increasingly central to drug discovery, yet they sit between two governance regimes. AI frameworks presume stable, well-characterized inputs, while organoid ethics focuses on donor consent, moral status, and tissue use, leaving downstream computational uses of organoid-derived data underregulated. This article argues that the convergence of living biological variability with algorithmic decision-making creates a governance vacuum in high-stakes preclinical contexts. The organoid-AI case is distinctive because two mature governance traditions, built on incompatible assumptions, intersect at a decision-critical point, and because organoid-derived biological states are transient and often irrecoverable. The article maps the structural causes of this gap, illustrates its practical consequences through three failure scenarios assessing likely probability and impact, and proposes an integrated two-pillar framework: graduated AI validation indexed to organoid functional complexity and tamper-resistant audit trails linking biological and algorithmic metadata. It concludes that proactive governance is morally preferable to delayed, crisis-driven regulation.
