Assessing the risks posed by chemicals is a difficult and imperfect business that reaches well beyond the scientific realm. There are many factors involved in balancing cost and benefit of such decisions that include science, governments, industry and the public.

Our recent feature looked at the ongoing debate around the risks of glyphosate – a herbicide that is one of the most widely used agrochemicals in the world. Its use has become increasingly controversial, particularly since the World Health Organization’s International Agency for Research on Cancer found in 2015 that it is ‘probably carcinogenic to humans’. Thousands of farmers have also been suing its developer, Bayer–Monsanto, claiming the product is responsible for their cancers, with victories for both sides in a long-running series of legal cases. But in its periodic review of the chemical, the European Chemicals Agency (Echa) decided to renew its approval in 2023, within some limits, ruling that the product is safe and that there is no scientific basis for a ban. It will be reviewed again in 2033.

The case illustrates how assessing risk is a complicated and nuanced process. The difference in opinion between IARC and Echa, for example, stems in part from the data sources used to reach their assessments – IARC uses only publicly available data, while Echa includes data from manufacturers – but it also reflects Echa’s evaluation of the practical reality and whether levels of real-world exposure are likely to be harmful. Then there are effects beyond human toxicology – for example, the evidence of harm to insect populations. Public opinion is also a key component that puts pressure on policymakers – once public sentiment has soured it can be difficult, if not impossible, to reverse.

Glyphosate joins a long list of materials that have made the journey from utility to ubiquity and then notoriety. Despite improvements in chemicals safety and risk assessment regulation over many decades, there often follows a similar pattern when chemicals leave the lab and market forces take over. This is when indiscriminate use or simply over-reliance can develop because a technology becomes dominant, and glyphosate is certainly a victim of its success here. PFAS is another recent example – an immensely useful class of chemicals that has become a synonym for environmental harm and whose utility might now be lost to us as a result. In the right place it is a material without peer, but regulators now faced with the challenge of responding to the harm posed by its overuse are implementing outright bans.

More responsive risk assessment approaches and better tools for toxicology, such as predictive models that can highlight whether a substance poses a risk, will help to make this process more effective. And so will chemists: their training includes understanding the risks of their work and these days many undergraduate courses also explicitly address the need to improve the sustainability of their practice – it’s now part of the Royal Society of Chemistry’s accreditation criteria for university courses.

What’s perhaps lacking in the system at the moment, however, is support or incentives to innovate and develop replacements. Switching to alternative pest control strategies or products, for example, also comes with significant risk – not least the costs of investment, along with the burden of proving they are safe and effective. There can also be an unwillingness to engage with debates around particular substances if those who rely on them feel there is no better option. If we hope to progressively substitute chemicals with safer alternatives, we need a system that enables innovation to find those alternatives and helps deliver them to market. We should never find ourselves in the position that a product becomes too big to fail.