Quick hits: April 2024

The volume of disinformation generated by artificial intelligence, particularly deepfake images related to elections, has increased by an average of 130% per month on X in the past year.

These are some of the findings from a study developed and published by the Center for Countering Digital Hate (Ccdh), a British nonprofit organization committed to fighting online hate speech. To measure the phenomenon’s increase - the latest false photos involve Trump with African Americans, generated by his supporters - the study examined the four most popular image generators: Midjourney, OpenAI’s DALL-E 3, Stability AI’s DreamStudio, or Microsoft’s Image Creator. Specifically, the 130% increase on X stemmed from the number of comments on the platform’s Community Notes, the tool through which users make observations on false and misleading posts. All the companies examined, by the way, have explicitly stated Google will begin penalizing content generated by artificial intelligence created solely to rank high in its search results.

This is a change that could have a ripple effect on the quality of what people see online.

As explained by the company in a post on its official blog, the novelty stems from improvements to the algorithms it uses to rank websites and the content that populates them. Specifically, the changes, effective by March, will affect three types of content, or “abuses,”