AI is being deployed in practical ways for suicide prevention, from flagging social media posts for the police to powering hybrid support platforms. However, these positive examples exist alongside serious, documented failures that have cost lives. This underscores a universal finding from the research: the most effective and safest AI applications are those that integrate robust human oversight, rather than attempting to replace human judgement entirely. The talk will also provide examples of how AI is used in a practical, hands-on way in the “Open Up” Online Counselling Platform, including:

(i) Identifying Warning Signs: The platform uses natural language processing to analyse chat messages in real time, flagging words or phrases that may indicate suicidal thoughts or distress.

(ii) Prioritising High-Risk Cases: By comparing messages to a constantly updated watchlist of keywords, the AI system helps counsellors prioritise the most urgent cases, ensuring timely intervention.

(iii) Improving Service Efficiency: The use of big data analytics and AI helps the team better allocate its resources and optimise the counselling service.