'As an independent investigative journalist, it is my duty to serve the broad public interest' (CBS News/YouTube)

February 6, 2024   7 mins


The below forms part of the testimony Lee Fang delivered to the House Select Committee on Weaponization on 6 February, 2024. 


Chairman Jordan, Ranking Member Plaskett, esteemed members of the Select Committee,

As an independent investigative journalist, it is my duty to serve the broad public interest, to watch over governmental and corporate conduct, to safeguard the principles underpinning a free society, and to spotlight wrongdoing — whether it comes from the Left, the Right, or the centre. It is a privilege to present my work to this committee. Over the past 15 years, I have uncovered a variety of different forms of surveillance and censorship, documenting how special interest groups, politicians, and law enforcement agencies use these tactics to silence or weaken their enemies.

I am committed to reporting on surveillance and censorship, because I regard free speech as the bedrock of our republic and the linchpin of our democratic process. These principles are the defining characteristics that set America apart on the global stage. But freedom of speech has not endured without a fight. In every generation, political and corporate actors from across the ideological spectrum have seized on crisis and moral panics to demand restrictions on speech.

Now as ever, we must remain vigilant against efforts to undermine free speech and free expression. Recent technological advancements have created new opportunities to restrict speech without public input or accountability. The rapid development of artificial intelligence tools, in particular, offers powerful entities the unprecedented ability to monitor, flag, and censor billions of individuals at a scale and scope never before conceivable. While this technology holds significant benefits, it also carries substantial risks. Private security firms are increasingly marketing AI solutions to both government agencies and corporations. Mounting evidence suggests that state and business interests are already deploying this technology with the goal of stifling lawful discourse and silencing dissent.

Some of you may be familiar with my October 2022 investigation delving into the history of the US Department of Homeland Security’s Cybersecurity and Infrastructure Agency (CISA) and the FBI’s concurrent expansion into the policing of social media. Using court documents and evidence provided by a DHS whistleblower, I reported that government programs initially designed to curb foreign influence and incitement to terrorism had transformed into a broader campaign to suppress ordinary domestic speech. CISA’s expansive focus eventually touched on a wide range of political topics, from the 2020 presidential election to the origins of Covid-19 to criticism of the Ukraine-Russia war.

Two months later, in December 2022, I reported on a cache of Twitter’s internal corporate documents that became known as the Twitter Files. I gained access to internal emails, tools, and chats that confirmed my earlier reporting about CISA. I revealed that law enforcement agencies like CISA and non-governmental partner organisations consistently pressured Twitter executives to censor political speech under the guise of combating misinformation. I have since published many articles on my Substack based on these documents, including a piece released on Monday.

In my latest report, I reveal that a non-profit anti-misinformation election consortium provided an inaccurate tip to CISA in November 2020 about a New York Times journalist’s observations about delays in the presidential vote count in Green Bay, Wisconsin. Instead of verifying the validity of the tip, CISA hastily produced a “Misinformation Report” about the incident, arguing that the tweet had promoted falsehoods about the election. The agency then lobbied Twitter to restrict access to the reporter’s tweet. In response, Twitter shadowbanned the Times reporter’s tweet, effectively rendering it invisible to most users — despite substantial evidence supporting the journalist’s accuracy.

I understand the stated intent behind these efforts — the desire to uphold the highest standards for our elections and deter any illicit manipulation. However, as this and many other cases illustrate, the government often errs and acts in a politically motivated manner. For this reason alone, it is inappropriate for law enforcement or intelligence agencies to act as the arbiter of permissible journalism. Moreover, government censorship of truthful and accurate speech — rather than dispelling conspiracy theories — serves only to exacerbate the erosion of public trust in our elections.

My work on the Twitter Files and other reporting on this topic differs slightly from recent testimony heard by this committee. I have also shed light on private entities’ attempts to control and curtail public discourse on major areas of public policy. I revealed that a group called Public Good Projects regularly collaborated with Twitter during the pandemic to censor specific social media accounts because of their criticism of establishment views around Covid-19 vaccines, while amplifying accounts supportive of vaccines and government viewpoints. This particular censorship campaign was entirely funded by biopharma lobbyists that represent Moderna and Pfizer.

In some instances, the censorship demands targeted users who simply expressed opposition to “vaccine passports” – the proposal to make a person’s right to travel and engage in certain forms of commerce dependent on their vaccination status. The validity of the idea of “vaccine passports” is a reasonable topic of debate in a free society, and should not be repressed by a drug company with financial interests at stake.

More recently, in collaboration with UnHerd and RealClearInvestigations, and with co-author Jack Poulson, I reported that Moderna relaunched efforts to influence vaccine discourse, again working with Public Good Projects. Moderna also employed the services of the artificial intelligence firm Talkwalker to monitor vaccine-related conversations across 150 million websites, including social media and gaming platforms like Steam. Moderna labelled almost any critic of the company or vaccine policies as “high risk” for spreading misinformation, including reputable academics like Stanford University’s Jay Bhattacharya and local elected officials who simply voiced objections to vaccine mandates.

The company even targeted tennis champion Novak Djokovic in its misinformation reports, not for any statement he made but for his accomplishments. In one report, Moderna warned that: “The optics of Djokovic, whose vaccine opposition barred him from competing in the 2022 US Open, returning to and winning the Moderna-sponsored [2023 US Open] competition bolsters anti-vaccine claims that vaccines — and mandates — are unnecessary.” In other words, Djokovic’s alleged act of “misinformation” was his unvaccinated athletic ability. There are many other examples in my reporting, beyond the Twitter Files and Moderna documents, that show government and corporate interests trying to stifle free speech on public safety and national security grounds.

Last month, I reported on confidential documents chronicling the activities of Logically, a British artificial intelligence firm that has expanded into the US market with the guidance of former DHS intelligence director Brian Murphy. This individual was previously responsible for the illicit surveillance of Left-wing protesters and the creation of dossiers on journalists critical of the government. Logically has quietly secured US military contracts and piloted a program for the Chicago Police Department to employ its AI platform to analyse local hip-hop music to predict crime. The state of Oregon also retained Logically to combat so-called election misinformation during the 2022 midterm elections. The company is now competing for contracts to monitor and remove alleged social media misinformation in the upcoming 2024 presidential election.

It is important to underscore why the American public should be aware of the British AI firm Logically. The United Kingdom government awarded Logically multi-million dollar contracts to combat misinformation about the Covid-19 pandemic. The company instead surveilled activists and academics who expressed legitimate forms of speech, including thoughtful concerns about pandemic lockdowns and vaccine passports, according to a recent watchdog report on the firm’s activities. Logically previously boasted of a special partnership with Meta, the parent company of Facebook and Instagram, to automatically suppress and label content they deemed as misinformation, giving the company immense influence over content moderation decisions. It is another case of the government outsourcing censorship to a private entity with a record of targeting lawful speech around hotly contested political issues.

I want to emphasise that this focus on corporate influence and government surveillance is not a recent development in my work. In 2015, I uncovered evidence of agricultural industry lobbyists ghostwriting state legislation designed to criminalise the recording of slaughterhouses and dairy facilities. The legislation was designed to imprison journalists and animal rights activists who document conditions in slaughterhouses. Many of these “Ag-Gag” laws have since been overturned by the courts for violating the First Amendment. I have also previously reported on government and private sector surveillance and censorship targeting environmental groups, Black Lives Matter activists, and labour unions.

In December, I exposed the efforts of pro-Israel activists to stifle pro-Palestinian voices. One notable example, relevant to today’s hearing, involves the deployment of artificial intelligence bots to mass-report social media posts expressing solidarity or sympathy for Palestinians as violations of platform terms of service. Many of the targets appear to be individuals using innocuous hashtags like #FreePalestine, unrelated to hate or violence. There are over 1,000 documented cases of Meta removing peaceful content advocating for Palestinian human rights. The use of artificial intelligence bots to suppress nonviolent and benign forms of political expression during a conflict illustrates another case of organised censorship.

Last summer, I profiled the activities of Flashpoint, among the cottage industry of so-called threat intelligence firms that use artificial intelligence to monitor activist behaviour. The company began by infiltrating al-Qaeda chat rooms and selling that data to the FBI. Critics argue that the firm played a central role in the wrongful convictions of several Muslim Americans. Flashpoint, which still maintains ties with the FBI, has more recently infiltrated the private communications of ordinary airline workers who organised against coercive and ultimately unconstitutional vaccine mandates. Many of these anti-mandate activists were conservative Trump supporters.

I present these examples to underscore how censorship affects dissenting voices of all ideological stripes. I understand that in our intensely polarised environment, free speech has become a divisive issue — often misused by politicians seeking a scapegoat for societal ills. But free speech is critical for activists spanning the ideological spectrum, and to every journalist committed to uncovering the truth. And history teaches us that when governments and private entities demand new censorship authority to attack dissidents of a particular stripe in one era, those tools are just as likely to be used against an entirely different set of actors a few years later. Today’s cheerleaders for an unaccountable “content moderation” regime may very well be tomorrow’s victims of that same system.

Before I close, I wish to draw your attention to a hearing held nearly 12 years ago by another House committee, addressing a remarkably similar topic. The committee convened a hearing to demand answers from the Department of Homeland Security about revelations that the agency had retained General Dynamics, a defence contractor, to conduct round-the-clock monitoring of social media.

Republicans controlled the House at the time, but members of the committee from both sides of the aisle raised concerns about DHS practices. Then-Rep. Jackie Speier, a Democrat from California, called for assurances that DHS was not infringing upon “civil rights and civil liberties of those who choose to use social media” or “spying on lawful activities”. Then-Rep. Patrick Meehan, a Republican from Pennsylvania, cautioned against DHS and its contractors monitoring private citizens’ speech, which he said “could have a chilling effect on individuals’ privacy rights and people’s freedom of speech and dissent against their government”.

That hearing looks quaint today. In terms of bureaucratic and technological capabilities, the General Dynamics scandal was a relic of another era. It occurred prior to the emergence of CISA, the DHS agency with a multi-billion-dollar budget, actively monitoring Americans for wrong speak on contentious issues. It transpired before significant advancements in AI allowed for far more invasive monitoring of constitutionally protected speech through sentiment analysis. And it pre-dated the creation of numerous anti-misinformation NGOs and private-sector firms competing for contracts to control and censor the internet.

Nonetheless, the 2012 hearing serves as a reminder that concerns about censorship and surveillance need not break down along partisan lines. Lawmakers from both sides of the aisle can still unite to voice concerns about the erosion of free expression and government efforts to regulate political discourse. I implore this committee to rise above partisanship and treat the threat posed by online surveillance and censorship as an American issue, affecting all of us equally. I am deeply grateful for the opportunity to testify today, and I eagerly await your questions. Thank you.

Lee Fang is an investigative journalist and Contributing Editor at UnHerd. Read his Substack here.