The National Science Foundation (NSF) is awarding millions of dollars in grants to universities and companies to develop AI-based technologies that contribute to the Biden Administration's campaign to combat misinformation.
According to some experts and media outlets, one of the tools - called Convergence Accelerator Track F - could quickly manipulate, "identify and censor U.S. dissent on issues such as vaccine safety and electoral integrity." According to Mike Benz, executive director of the Foundation for Freedom Online:
This is really an embodiment of the censorship framework of the whole society. We see it now in full swing. What has been seen is a grafting of these concepts of misinformation and disinformation that escalated to such high levels of intensity that they became a formal and tangible government program to fund and accelerate censorship science.
A program similar to the one used against ISIS
Benz stated in a report that these new technologies have a strong similarity with those created in 2011 by the Defense Advanced Research Projects Agency (DARPA), in its program of Social Media in Strategic Communication (SMISC), in order to counteract the disinformation from the Islamic State (ISIS) in the uprising of the so-called Arab Spring in the Middle East more than a decade ago.
The initial idea of these technologies was to track terrorists - who wanted to attack the U.S. - by studying their social media posts. This program helped "to identify disinformation and deception campaigns and counteract them with truthful information." It had four specific objectives:
- Detect, classify, measure and track (a) the formation, development, and dissemination of ideas and concepts (memes); and,
(b) erroneous, intentional or misleading messages and information.
- Recognize the structures of persuasion campaigns and influence operations in communities and social networking sites.
- Identify participants and intent, and measure the effects of persuasion campaigns.
- Counter messages detected with adversarial influence operations.
Benz claimed that "DARPA has been funding an AI network using the science of social network mapping going back to at least 2011-2012" and pointed out:
One of the most troubling aspects of the Convergence Accelerator Track F domestic censorship projects is how similar they are to military-grade social network monitoring and censorship tools developed by the Pentagon for overseas counterinsurgency and counterterrorism contexts.
NSF, with an annual budget of $10 billion, asked Congress for an 18.7% increase in appropriations in its latest budget request.
Tool aimed at those who question vaccines or elections.
The report noted that the new version of this technology is aimed at people who are wary of the adverse effects of the covid-19 vaccine and skeptics of the election results in recent years:
The Convergence Accelerator Track F was created by the Trump Administration to address grand challenges such as quantum technology. When the Biden Administration came to power, they basically took this infrastructure for multidisciplinary purposes. They are funding artificial intelligence, censorship capabilities, to censor people who distrust the government or the media.
The advent of covid-19 led to "normalizing censorship in the name of public health," Benz noted, "and then, in the run-up to the 2020 election, all kinds of political censorship was shoehorned in as acceptable."
The "threat" of social networks
Just the News contacted DARPA for more information about the program, the agency justified itself on the grounds that the program "succeeded in developing a new science of social network analysis to reduce adversaries' ability to manipulate local populations outside the U.S.":
Given the importance of the threat posed by adversarial activities on social networking platforms, we are working to make many of the technologies under development open and available to researchers in this space.
Mike Pozmantier and Douglas Maughan, director and office head of NSF's Convergence Accelerator Track F program, respectively, did not respond to requests for comment.