Skip to main content
Parents   »   AI technology prevent toxic language

AI technology prevent toxic language

The Varannan Vecka-app uses AI technology to prevent toxic language between parents. The goal is to secure parents’ communication and help reduce long-term conflicts that risk harming children.


What is toxic language?

Most people have probably heard of ‘online hate’, ‘hate speech’ or ‘toxic language’. But what does it look like between parents? Examples can be abusive slurs (idiot, whore etc) or profanity (bastard, fuck). It can also be derogatory descriptions like “you are useless and stupid”. It can also be about offensive emojis, negative irony or other negative messages.

Psychological effects of toxic language

Those who experience toxic language and hate speech repeatedly often feel very bad psychologically. They can loose energy, become withdrawn and/or depressed. A verbally abused parent may have to focus more energy on dealing with toxic communication than being a strong and happy parent. Therefore, verbal abuse can have a direct effect on the children. They can become lonely when parents are stuck in serious conflicts.

Effects of verbal abuse between parents

In practice, recurrent verbal abuse almost always lead to a fast deterioration in parental cooperation. The victim needs to protect themselves and often cuts down on communication. The offender then often feels even more mistreated. Maybe they continue with their destructive behavior which can escalate.

For children, this is often a very difficult situation. Parents are often stressed, feel bad before and during handovers, may not greet each other. In worst case they start arguing in front of the children. Since these are conflicts between adults, children are powerless and can do nothing to change the situation. This often leads to children feeling sad, afraid and/or angry. It can lead to serious physical and emotional stress over time.

How TalkSafe can improve communication

The TalkSafe-technology can be described as an AI-filter that detects toxic language while it is being written. If a parent starts writing a chat message with toxic language the AI filter warns the parents making the message yellow, orange or red depending on the degree of toxicity. The message can still be sent, but the sender has received a clear warning that the message to the co-parent is most likely toxic and therefore like to spark or add conflict.

Under ideal conditions when users express themselves in a similar way as TalkSafe is trained to identify, correct warnings go out in 90 percent of the cases. This means that nine out of ten messages containing toxic language are flagged before the parent presses send. The hope is that the technology will give the parent a micro-impulse to stop, think and choose other words that are less likely to spark conflict. When there is less or no toxic language in the communication building a solid co-parenting team is possible.

The technology behind SafeTalk is the result of a collaboration between researchers at Uppsala University, Stockholm University and the Mind Intelligence Lab.

TalkSafe is part of the chat and is included in the total price of SEK 89 / month for the entire VarannanVecka-family (no additional costs for users).

Summary

  • Toxic language can be, for example, profanity, offensive words or derogatory messages, negative emojis or negative irony.
  • Those who are repeatedly subjected to toxic language often feel bad mentally and often withdraw from contact.
  • TalkSafe is unique AI technology which work as an AI filter that detects and warn the writer if a message contains toxic language.
  • The goal with TalkSafe is for the sender to stop, think and choose other words that do not drive conflict.
  • TalkSafe is part of the Varannan Vecka-chat and does not add any extra costs for the users, ie SEK 89 / month for the entire varannavecka-family.

If you have questions about TalkSafe or need help from our experts, please e-mail elisabeth@varannavecka.app. We love talking to our users!

Elisabeth Scholander