Generative AI models are susceptible to a kind of cyberattack called “data poisoning,” whereby malicious actors intentionally manipulate known source material to change the model’s understanding of an issue. It’s like a high-tech version of giving a school rival a fake exam answer key.
Researchers say that concerns of data poisoning are mostly hypothetical at this point, but showed in a new report how Wikipedia entries could be edited at strategic times to ensure the incorrect information is captured by models scraping the online encyclopedia. It’s an early warning to AI companies and those who depend on it that attackers could soon find creative ways to target the most powerful models and exploit vulnerabilities.
Data poisoning isn’t all bad: Some copyright holders are using a form of data poisoning as a defensive mechanism to prevent AI models from gobbling up their creative works. One program called Nightshade was developed to distort an image when it’s ingested by a large language model.
For Gen Z, changing jobs can mean a bigger paycheck. Younger workers continue to see the largest gains from switching, offering a glimpse into how career mobility is evolving across generations. Subscribe to Bank of America Institute to explore the data.
Tens of thousands still remain missing after the RSF’s siege of El Fasher. The RSF also remains in control of the city, and the civil war continues to drag on with seemingly no end in sight.
After weeks of student-led protests over poor conditions in French schools, the government has promised to replace its controversial university admissions platform by 2028.
In this Quick Take, Ian Bremmer breaks down Brazilian President Luiz Inácio Lula da Silva’s historic re-election victory and what it means for the country’s political future and democracy.