**FDA January 2026 CDS guidance: enforcement discretion widened – AI decision-support now reaches clinicians without premarket review**
In a development that heightens exposure for healthcare and financial services organisations in Canada and the UK, the US Food and Drug Administration (FDA) has announced its intention to exercise more enforcement discretion when it comes to reviewing Artificial Intelligence (AI) decision-support systems. This means that AI technologies can now be deployed without premarket review, which may lead to an increased risk of unexamined AI decisions reaching clinicians.
This development directly affects organisations buying or deploying AI in regulated sectors, as they may not have the same level of control over the technology and its claims. The widening of enforcement discretion increases the likelihood that organisations will need to rely on internal teams or vendors for validation, which can lead to a heightened risk of biased decision-making processes.
Organisations are exposed to capability risk – do AI solutions really perform as claimed? – governance risk – can the organisation defend AI decisions when regulators ask questions? – and liability risk – what is an organisation accountable for when the AI makes a mistake?
To mitigate these risks, prudent organisations should take immediate steps to assess the capabilities of their AI solution against real-world operations. This includes evaluating AI claims against internal data sets, assessing human oversight mechanisms, and identifying potential biases in decision-making processes.
Straven & Co examines AI decisions before they are acted on by providing independent validation. The firm tests proposed solutions against a client’s real operations, people, governance obligations, and legal exposure, delivering a plain verdict – proceed, proceed with conditions, or do not – to protect clients from the costs of unexamined AI decisions.