- Trump wants to force the sale of
TikTok to a US tech company in order to keep Americans' data out of the hands of the Chinese government. - But the app's most likely buyers,
Microsoft and Oracle, play a massive role in helping the US government spy on its own citizens. - The companies provide a range of technologies to law enforcement agencies, including facial recognition software, cloud computing services, and "predictive policing" algorithms — many of which rely on social media data.
- Amid mounting evidence that these tools can exacerbate racial disparities and growing pressure on tech companies to reexamine their relationships with US law enforcement, the potential sale of TikTok raises questions about Microsoft and Oracle's roles in America's own surveillance apparatus.
The future of TikTok, and the data of its more than 100 million American users, now hangs in the balance as President Donald Trump seeks to force its sale to an American tech company.
The reason? Trump claimed in two executive orders targeting TikTok's parent company, ByteDance, that the Chinese government could force the Beijing-based firm to share user data with its intelligence agencies so it can spy on Americans.
Trump isn't alone when it comes to concerns about giving China a backdoor into the phones of US diplomats, military officials, or politicians. Democratic presidential nominee Joe Biden's campaign, the Democratic and Republican parties, various federal agencies, and nearly every branch of the US military have all banned its use, also citing national security.
The president has gone a step further, however, arguing that everyday Americans are equally at risk and a US firm would be a better steward over their data, presumably keeping them safer as a result.
But as America faces a reckoning with racially biased police, policing practices, and policing technology, the tech industry's cozy relationship with law enforcement is being called into question. In many cases, those companies are supercharging the same authorities that have been accused of targeting Black and indigenous people, immigrants, Arab-Americans, and other groups that have borne the brunt of state violence.
Beijing may be able to blackmail a corporate executive or high-ranking government official, but for many Americans, tech-enabled police forces pose a far more imminent danger, and TikTok's most likely buyers have a strong track record of cooperating closely with law enforcement — sometimes by analyzing the very social media data they could soon be tasked with safeguarding.
Government ties
Microsoft and Oracle have stood as the frontrunners to buy TikTok largely because of their history of cooperation with law enforcement. In a press conference earlier this month, Trump specifically cited Microsoft's existing "high-level" security clearances from its work with the US government.
Microsoft and Oracle both declined to comment for this story.
Both companies have a long history providing technology solutions to US military and law enforcement agencies of various stripes. Oracle got its start in 1977 when the CIA contracted cofounder Larry Ellison to build it a database. Microsoft started building the New York Police Department's notorious surveillance network in 2009.
They've also been reliable partners to their government counterparts even when that involved legally or ethically dubious work. Microsoft gave the NSA access to its users' encrypted messages and Skype video calls as part of the notorious PRISM program unearthed by Edward Snowden in 2013. Ellison defended that same NSA surveillance dragnet, calling it "essential."
Eleven years earlier, Ellison argued in a 2002 New York Times op-ed that the US should create "a national security database combined with biometrics, thumb prints, hand prints, iris scans."
As of the present day, Oracle and Microsoft have developed and provided exponentially more powerful surveillance technology to law enforcement.
In a July report, The Intercept's Michael Kwet provided a detailed look into the suite of Microsoft services deployed in the name of public safety. Some notable ones include:
- The NYPD surveillance network mentioned earlier, which in just four years amassed two billion license plate scans, 15 million police reports, 33 billion public records, and footage from 9,000 surveillance and 20,000 body cameras — all of which feed into a "predictive policing" algorithm to identify where future crimes might occur.
- The Microsoft Advanced Patrol Platform, an internet-of-things enabled patrol car that syncs various data sources to offer real-time intel to officers, and which the company hopes to pair with AI-powered "pattern matching" to help them identify "bad actors."
- AI
analysis of police surveillance data via its Azure cloud computing service, which cities like New York, Atlanta, Chicago, and Detroit use.
Oracle, through its "Social-Enabled Policing" tools, aims to "leverage the power of social media" to help law enforcement with "the prevention, detection and solving of crime and disorder." Its flagship product, an AI-powered data analysis platform, collects more than 700 million messages daily from over 40 million social media websites.
Microsoft and Oracle have largely flown under the radar compared to companies like Palantir that market law enforcement partnerships more explicitly. Because Microsoft and Oracle primarily sell software that can also be tailored for non-law enforcement uses, and work with many third party partners and contractors, identifying whether or how extensively they're involved with a given agency's IT infrastructure can be difficult.
Scrutinizing surveillance
In recent years, surveillance technologies and the manner in which agencies use them have come under more intense scrutiny for potential civil rights and civil liberties abuses.
Facial recognition technology, predictive policing algorithms, and other AI-powered tools have frequently been shown to reinforce racial, gender, and other biases, and even led to a wrongful arrest in Michigan earlier this year.
The FBI's use of an internet surveillance program was ruled unconstitutional; US Customs and Border Patrol and other agencies have paid private companies for people's personal data to avoid obtaining warrants; and police departments have used everything from social media posts to Ring doorbell cameras to spy on and profile Black Lives Matter protesters in ways that evoke the FBI's controversial COINTELPRO program targeting Black activists from the 1950s to 1970s.
Following George Floyd's killing, and amid aggressive and at times violent responses to protesters by law enforcement in Seattle (near Microsoft's headquarters), more than 250 Microsoft employees sent a letter to the company's executives asking Microsoft to end its contracts with police departments.
Shortly after, Microsoft said it wouldn't sell facial recognition technology to police until stronger civil rights protections are in place, but the company's commitment left major loopholes for its other services that enable police to use facial recognition and other surveillance technologies.
Yet even aside from providing IT infrastructure, Microsoft has assisted law enforcement by consistently complying with requests for user data, agreeing to such requests in 80% to 90% of cases since it began publishing data in 2013. (Oracle does not publish comparable reports).
Both companies refused to answer questions about how they currently handle government requests for data, how they ensure lawful uses of their technologies or handle cases where law enforcement agencies misuse the tools, and how they would protect US TikTok user data if they acquired the app.
With TikTok's rise as a platform for anti-racism activists and law enforcement's increased use of social media to surveil protesters in the US, some Americans may have more to fear with their data in the hands of Microsoft or Oracle than they do with China's ByteDance.