In the report called "Freedom on the Net 2023: The Repression Power of Artificial Intelligence", the organization indicated that before the second round of the 2022 presidential elections in Brazil, a 'deepfake' video – digital manipulation of facial appearance – spread across the Internet falsely portraying a reliable media outlet reporting that the now former President Jair Bolsonaro was leading the polls.
This is just one of the report's examples of leveraging Artificial Intelligence (AI) to intensify disinformation campaigns on the Internet.
The report also gave as an example the company Mind Force, based in Israel, which operated a network of accounts in Angola and published content in support of the ruling party, the Popular Movement for the Liberation of Angola, and against the country's main opposition party. country.
"A Mind Force employee publicly disclosed that the Government was a client," says Freedom House.
Between June 2022 and May 2023, the period in which the organization carried out the survey, at least 47 Governments – such as Russia, Brazil, China, India, Angola, Mexico, Venezuela or Nigeria – used commentators to manipulate online discussions to their advantage, double the number recorded a decade ago.
Meanwhile, AI-based tools that can generate text, audio and images have quickly become more sophisticated, accessible and easier to use, fueling a worrying rise in these disinformation tactics.
During the period under review, the new technology was used in at least 16 countries to sow doubts, defame opponents or influence public debate, including China, France, Brazil, the United Kingdom, the United States or Canada, according to the non-profit organization Washington-based Freedom House.
However, the report finds that while innovations in AI have contributed to the 13th consecutive year of global decline in internet freedom, older forms of digital repression have continued to proliferate.
Iran suffered its worst score drop of the year as authorities shut down internet services and blocked social media to quell anti-government protests.
Countries such as China, Cuba, Indonesia, Iran, Jordan, Myanmar (formerly Burma) or Uzbekistan are among the territories whose Governments most apply traditional forms of online censorship.
Brazil was the only Portuguese-speaking country analyzed in this segment, with the report pointing out that it blocked websites and forced the removal of online content.
In relation to global freedom on the Internet, which decreased for the 13th consecutive year, the report classified Brazil and Angola, the only Portuguese-speaking countries analyzed, as "partially free".
This category also includes countries such as Ukraine, Hungary, Mexico, India or Morocco.
In the total of 70 countries covered by the report, Freedom House identified 17 states as "free" – including France, the United Kingdom, Germany, Italy, the United States or Canada –, 32 as "partially free" and 21 as "not free" – such as Russia, China, Iran, Myanmar, Venezuela or Cuba.
Freedom House indicated that as conditions deteriorated in 29 countries, they improved in another 20, despite attacks on freedom of expression becoming more common around the world.
"Freedom on the Net" is an annual study of human rights in the digital sphere. The project assesses Internet freedom in 70 countries, representing 88 percent of users worldwide. This report, the 13th in its series, covered developments between June 2022 and May 2023.
More than 85 analysts and consultants contributed to this year's edition, using a standard methodology to determine each country's Internet freedom score on a 100-point scale, with 21 separate indicators relating to barriers to access, content limits and violations of user rights.
