2,5 suicides are committed every hour in Spain. When they are alone facing their fears, young people use to show various signs of their condition on social media. Sadly, sometimes we can´t identify these red lights on time.

What if new technologies could be able to be part of the solution by interpreting these signs?

Code of Hope is an Artificial Intelligence able to detect suicidal behaviors on social media, by analyzing public posts and its emotions. More than just search for keywords or tags, this advanced A.I. translates data into how an individual is actually feeling.

Based on Robert Pluthick´s theory, the algorithm was developed to create an emotional profile after analyzing 32 different emotions. Thus, a pattern is shaped regarding the data collected from suicidal profiles. This pattern is used by the AI and its machine learning to detect suicidal behaviors with 97% of assertiveness.

Once these profiles are detected on Twitter, a bot starts to follow them automatically. Just by changing the NGO profile name, a message is sent to them to show that they´re not alone.

Then a new Twitter service enters in action: the DM card, offering help on a virtual chat with volunteers from Telefono de La Esperanza.

With data collected from suicidal profiles in Social Networks, the A.I. created an emotional pattern. Once the algorithm understands the emotional pattern of depressive profiles, it learns to detect suicidal behaviors online.

More than just search for keywords or tags, this advanced A.I. translates this data into how an individual is actually feeling. If the profile is detected with a high-risk depression, a bot starts to follow this profile automatically. Then a new Twitter service starts to act: a virtual chat with volunteers from Telefono de La Esperanza, an NGO that works with depressive people over the phone for more than 50 years.

In between depression and suicide there´s loneliness. Code for Hope acts right on this crucial point. After detecting the depressive profiles, the AI connects them right away with a volunteer to talk to.

The algorithm of Teléfono de la Esperanza analyzes the last 100 tweets of different users and identifies profiles that are showing emotional alterations similar to suicidal behaviors.

Once the high-risk profile is identified, the modified profile name “You´re not alone, wanna chat?” starts to follow them automatically, using a bot.

Clicking on the profile, a new Twitter´s service, the DMCard, offers help via a virtual chat.

After the first interaction, the AI continues to analyze every identified profile looking for changes on their emotional profile. If their emotional status improves, those users are put in stand by. If not, the system creates a new report and sends it to Teléfono de la Esperanza.

More than 1 million profiles analyzed on Twitter in Spain.

+300 High-Risk profiles detected so far, 8-10 detected per week.

All of them are being followed, contacted and treated by volunteers at Telefono de La Esperanza.