Data poisoning, or the act of introducing corrupted data during the training phase of artificial intelligence systems, is an emerging trend in social networks. The manipulation of AI in social networks raises ethical concerns, especially regarding companies like OpenAI that collect large amounts of user data. The appropriation of personal and collective memories through AI-powered conversation tools has significant implications for how individuals construct their identities. Countermeasures like Nightshade, which alters the training data of AI models, have been developed to protect artists’ work from unauthorized use by AI companies. However, these countermeasures can also be misused by extremist groups for their own purposes. Overall, data poisoning highlights the need for ethical considerations and safeguards in the development and use of AI in social networks.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| Data poisoning in artificial intelligence | Manipulation of AI through corrupted data | Enhanced security measures to prevent data poisoning | Privacy concerns and desire for control over AI systems. |