[ad_1]
TECHNOLOGY
Pessimism about misinformation grows and social media owners are seen as the biggest threat
Academics from areas such as computer science, data engineering and social sciences and humanities from 66 countries participated in the research
Published on October 24, 2024 at 09:10
Social media Credit: Image: Viktollio | Shutterstock
Research by the International Panel on the Information Environment (IPIE) revealed worsening expectations among experts about disinformation. The study points to political leaders, governments and social network owners as the main threats to the quality of the global information environment and places artificial intelligence (AI) at the center of concerns. Analysts interviewed by Estadão consider that, although AI had a smaller impact than expected in the 2024 elections, institutions must remain alert for the next elections.
More than 400 academics from areas such as computer science, data engineering and social sciences and humanities, from 66 countries, were interviewed by IPIE. While more than half (54%) expected conditions to worsen in 2023, the number rose to 63% in 2024. In developing countries, such as Brazil, 75% said they were very pessimistic about the future of the information environment. One of the main reasons for this pessimism is the use of AI.
Social network owners were cited as the main threat to the information environment by 15.5%. Governments and politicians from the interviewees’ country of origin came next, with 12.1%. Then, governments and politicians from other countries and state media. The lowest risk, according to those interviewed, comes from journalists and the press, cited as a threat by just 1.5% of participants.
“The control that these entities (platforms) exercise over content distribution and moderation policies significantly affect the quality and integrity of information,” said IPIE President Phil Howard. “The unfettered power of these entities poses a serious risk to the health of our global information environment.”
For more than half of survey participants, artificially generated videos, voices, images and texts pose a risk. Data Privacy Brasil, which monitors the use of AI in partnership with Desinformante and Aláfia Lab, recorded 26 cases during the first round of the 2024 electoral process in the country. Deepfakes, chatbots and artificially created jingles were the main types of AI used .
Among the incidents, even “deepnudes” were recorded, fake images and videos with sexual content. Candidates defeated in the São Paulo dispute, Tabata Amaral (PSB) and Marina Helena (Novo) were victims. The list of fake videos also included an alleged hug between Pablo Marçal (PRTB) and Tabata.
“It is a permanent risk. Hence the importance of being observed and contained by institutional measures, especially the actions of the Justice”, said the director of Data Privacy Brasil, Rafael Zanatta “It is necessary to move forward from this claim to the next, reassessing how much the current rules They are adherents to a reality that is always changing.”
Head of Programs at IPIE, Egerton Neto stated that the next elections will be challenging. “The hardest is yet to come. The 2026 and 2028 elections will take place in an even worse information environment. This puts pressure on institutions both to regulate the use of AI and to explore ways in which AI can be an ally, as in deepfake detection.”
The information is from the newspaper O Estado de S. Paulo.
[ad_2]
Source link