I applaud Tim Trevan's call to reform lab biosafety, but disagree with his argument for using the nuclear industry as a model (Nature 527, 155–158; 2015).

Nuclear facilities are strictly regulated and ensure that potential hazards arising from process changes are engineered out (see go.nature.com/qyzoth). Yet scientists are not process-driven: being autonomous and creative, they need freedom to change and require a dynamic safety culture that can accommodate new challenges. These include the replacement of humans by technology, reduced supervision and declining safety competencies — none of which applies to the nuclear industry.

Chasing a 'zero harm' mantra can actually promote a poor safety culture because it is an outcome rather than a goal (for examples of alternative approaches, see go.nature.com/xgupio and go.nature.com/gcjqfl). As Trevan points out, an effective safety culture is measured through engagement, understanding and care for everyone's well-being.