Twitter Bots Are Spreading Massive Amounts of COVID-19 Misinformation

• Read original article here

Twitter Bots Are Spreading Massive Amounts of COVID-19 Misinformation
Share
Search:
Explore by topic
FOR THE TECHNOLOGY INSIDER
Topics
Follow IEEE Spectrum
Support IEEE Spectrum
IEEE Spectrum is the flagship publication of the IEEE — the world’s largest professional organization devoted to engineering and applied sciences. Our articles, podcasts, and infographics inform our readers about developments in technology, engineering, and science.
About IEEE Contact & Support Accessibility Nondiscrimination Policy Terms IEEE Privacy Policy
© Copyright 2022 IEEE — All rights reserved. A not-for-profit organization, IEEE is the world's largest technical professional organization dedicated to advancing technology for the benefit of humanity.
IEEE websites place cookies on your device to give you the best user experience. By using our websites, you agree to the placement of these cookies. To learn more, read our Privacy Policy.
Enjoy more free content and benefits by creating an account
Saving articles to read later requires an IEEE Spectrum account
The Institute content is only available for members
Downloading full PDF issues is exclusive for IEEE Members
Access to Spectrum's Digital Edition is exclusive for IEEE Members
Following topics is a feature exclusive for IEEE Members
Adding your response to an article requires an IEEE Spectrum account
Create an account to access more content and features on IEEE Spectrum, including the ability to save articles to read later, download Spectrum Collections, and participate in conversations with readers and editors. For more exclusive content and features, consider Joining IEEE .
Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, archives, PDF downloads, and other benefits. Learn more →
Close
Access Thousands of Articles — Completely Free
Create an account and get exclusive content and features: Save articles, download collections, and talk to tech insiders — all free! For full access and benefits, join IEEE as a paying member.
internet ieee covid19 resources Twitter bots misinformation covid-19 social media
Back in February, the World Health Organization called the flood of misinformation about the coronavirus flowing through the Internet a “massive infodemic.” Since then, the situation has not improved. While social media platforms have promised to detect and label posts that contain misleading information related to COVID-19, they haven’t stopped the surge.
But who is responsible for all those misleading posts? To help answer the question, researchers at Indiana University’s Observatory on Social Media used a tool of their own creation called BotometerLite that detects bots on Twitter. They first compiled a list of what they call “low-credibility domains” that have been spreading misinformation about COVID-19, then used their tool to determine how many bots were sharing links to this misinformation.
Their findings , which they presented at this year’s meeting of the Association for the Advancement of Artificial Intelligence , revealed that bots overwhelmingly spread misinformation about COVID-19 as opposed to accurate content. They also found that some of the bots were acting in “a coordinated fashion” to amplify misleading messages.
The scale of the misinformation problem on Twitter is alarming. The researchers found that overall, the number of tweets sharing misleading COVID-19 information was roughly equivalent to the number of tweets that linked to New York Times articles.
We talked with Kai-Cheng Yang , a PhD student who worked on this research, about the bot-detection game.
This conversation has been condensed and edited for clarity.
IEEE Spectrum: How much of the overall misinformation is being spread by bots?
Kai-Cheng Yang: For the links to the low-credibility domains, we find about 20 to 30 percent are shared by bots. The rest are likely shared by humans.
Spectrum: How much of this activity is bots sharing links themselves, and how much is them amplifying tweets that contain misinformation?
Yang: It’s a combination. We see some of the bots sharing the links directly and other bots are retweeting tweets containing those links, so they’re trying to interact with each other.
Spectrum: How do your Botometer and BotometerLite tools identify bots? What are they looking for?
Yang: Both Botometer and BotometerLite are implemented as supervised machine learning models. We first collect a group of Twitter accounts that are manually annotated as bots or humans. We extract their characteristics from their profiles (number of friends, number of followers, if using background image, etc), and we collect data on content, sentiment, social network, and temporal behaviors. We then train our machine learning models to learn how bots are different from humans in terms of these characteristics. The differences between Botometer and BotometerLite is that Botometer considers all these characteristics whereas BotometerLite only focuses on the profiles for efficiency.
Spectrum: The links these bots are sharing: Where do they lead?
Yang: We have compiled a list of 500 or so low-credibility domains. They’re mostly news sites, but we would characterize many of them as ‘fake news.’ We also consider extremely hyper-partisan websites as low-credibility.
Spectrum: Can you give a few examples of the kinds of COVID-related misinformation that appear on these sites?
Yang: Common themes include U.S. politics, status of the outbreak, and economic issues. A lot of the articles are not necessarily fake, but they can be hyper-partisan and misleading in some sense. We also see false information like: the virus is weaponized, or political leaders have already been vaccinated.
Spectrum: Did you look at whether the bots spreading misinformation have followers, and whether those followers are humans or other bots?
Yang: Examining the followers of Twitter accounts is much harder due the API rate limit, and we didn’t conducted such analysis this time.
Spectrum: In your paper, you write that some of the bots seem to be acting in a coordinated fashion. What does that mean?
Yang: We find that some of the accounts (not necessarily all bots) were sharing information from the same set of low-credibility websites. For two arbitrary accounts, this is very unlikely, yet we found some accounts doing so together. The most plausible explanation is that these accounts were coordinated to push the same information.
Spectrum: How do you detect bot networks?
Yang: I’m assuming you are referring to the network shown in the paper . For that, we simply extract the list of websites each account shares and then find the accounts that have very similar lists and consider them to be connected.
Spectrum: What do you think can be done to reduce the amount of misinformation we’re seeing on social media?
Yang: I think it has to be done by the platforms. They can do flagging, or if they know a source is low-credibility, maybe they can do something to reduce the exposure. Another thing we can do is improve the average person’s journalism literacy: Try to teach people that there might be those kinds of low-credibility sources or fake news online and to be careful. We have seen some recent studies indicating that if you tell the user what they’re seeing might be from low-credibility sources, they become much more sensitive to such things. They’re actually less likely to share those articles or links.
Spectrum: Why can’t Twitter prevent the creation and proliferation of bots?
Yang: My understanding is that when you try to make your tool or platform easy to use for real users, it opens doors for the bot creators at the same time. So there is a trade-off.
In fact, according to my own experience, recently Twitter started to ask the users to put in their phone numbers and perform more frequent two-step authentications and recaptcha checks. It’s quite annoying for me as a normal Twitter user, but I’m sure it makes it harder, though still possible, to create or control bots. I’m happy to see that Twitter has stepped up.
27 May 2022
11 min read
The airline and cellular-phone industries have been at loggerheads over the possibility that 5G transmissions from antennas such as this one, located at Los Angeles International Airport, could interfere with the radar altimeters used in aircraft.
Patrick T. Fallon/AFP/Getty Images
Blue
You’ve no doubt seen the scary headlines: Will 5G Cause Planes to Crash? They appeared late last year, after the U.S. Federal Aviation Administration warned that new 5G services from AT&T and Verizon might interfere with the radar altimeters that airplane pilots rely on to land safely. Not true, said AT&T and Verizon, with the backing of the U.S. Federal Communications Commission , which had authorized 5G. The altimeters are safe, they maintained. Air travelers didn’t know what to believe.
Another recent FCC decision had also created a controversy about public safety: okaying Wi-Fi devices in a 6-gigahertz frequency band long used by point-to-point microwave systems to carry safety-critical data. The microwave operators predicted that the Wi-Fi devices would disrupt their systems; the Wi-Fi interests insisted they would not. (As an attorney, I represented a microwave-industry group in the ensuing legal dispute.)
Whether a new radio-based service will interfere with existing services in the same slice of the spectrum seems like a straightforward physics problem. Usually, though, opposing parties’ technical analyses give different results. Disagreement among the engineers then opens the way for public safety to become just one among several competing interests. I’ve been in the thick of such arguments, so I wanted to share how these issues arise and how they are settled.
Battling for Bandwidth
Not all radio spectrum is created equal. Lower frequencies travel farther and propagate better through buildings and terrain. Higher frequencies offer the bandwidth to carry more data, and work well with smaller antennas. Every radio-based application has its own needs and its own spectral sweet spot.
Suitable spectrum for mobile data—4G, 5G, Wi-Fi, Bluetooth, many others—runs from a few hundred megahertz to a few gigahertz. Phones, tablets, laptops, smart speakers, Wi-Fi-enabled TVs and other appliances, Internet-of-things devices, lots of commercial and industrial gear—they all need these same frequencies.
The problem is that this region of spectrum has been fully occupied for decades. So when a new service like 5G appears, or an older one like Wi-Fi needs room to expand, the FCC has two options. For a licensed service like 5G, the FCC generally clears incumbent users from a range of frequencies—either repacking them into other frequencies nearby or relocating them to a different part of the spectrum—and then auctions the freed-up spectrum to providers of the new service. To accommodate an unlicensed service like Wi-Fi, the FCC overlays the new users onto the same frequencies as the incumbents, usually at lower power.
The FCC tries to write technical rules for the new or expanded service that will leave the incumbents mostly unaffected. It is commonplace for newcomers to complain that any interference they cause is not their fault, attributing it to inferior incumbent receivers that fail to screen out unwanted signals. This argument usually fails. The newcomer must deal with the spectrum and its occupants as it finds them. Strategies for accomplishing that task vary.
Alternative Realities
This radio tower, located near downtown Los Angeles, is bedecked with 6-GHz fixed-microwave antennas that serve area police and fire departments.
George Rose/Getty Images
Congress prohibits the FCC (and other federal agencies) from changing the regulatory ground rules without first soliciting and considering public input . On technical issues, that input comes mostly from the affected industries after the FCC outlines its tentative plans in a Notice of Proposed Rulemaking. There follows a back-and-forth exchange of written submissions posted to the FCC’s website, typically lasting a year or more.
Ordinarily, parties can also make in-person presentations to the FCC staff and the five commissioners, if they post summaries of what they say. Sometimes the staff uses these meetings to test possible compromises among the parties.
All this openness and transparency has a big exception: Other federal agencies, like the FAA, can and sometimes do submit comments to the FCC’s website, but they also have a back channel to deliver private communications.
The submissions in a spectrum proceeding generally make two kinds of points. First, the newcomers and the incumbents both present data to impress the FCC with their respective services’ widespread demand, importance to the economy, and utility in promoting education, safety, and other public benefits. Second, both the proponents and opponents of a new frequency usage submit engineering studies and simulations, sometimes running to hundreds of pages.
Predictably, the two parties’ studies come to opposite conclusions. The proponents show the new operations will have no harmful effect on incumbents, while the incumbents demonstrate that they will suffer devastating interference. Each party responds with point-by-point critiques of the other side’s studies and may carry out counter-studies for further proof the other side is wrong.
How do such alternative realities arise? It’s not because they are based on different versions of Maxwell’s equations. The two sides’ studies usually disagree because they start with differing assumptions about the newcomer's transmitter characteristics, the incumbent's receiver characteristics, and the geometries and propagation that govern interaction between the two. Small changes to some of these factors can produce large changes in the results.
Rather than settle anything, experiments just add fuel to the controversy.
Sometimes the parties, the FCC, or another government agency may conduct hardware tests in the lab or in the field to assess the degree of interference and its effects. Rather than settle anything, though, these experiments just add fuel to the controversy. Parties disagree on whether the test set-up was realistic, whether the data were analyzed correctly, and what the results imply for real-world operations.
When, for example, aviation interests ran tests that found 5G transmissions caused interference to radio altimeters, wireless carriers vigorously challenged their results. In contrast, there was no testing in the 6-GHz Wi-Fi proceeding, where the disagreements turned on theoretical analyses and simulations.
Further complicating matters, the disputed studies and tests do not predict interference as a binary yes/no but as differing probabilities for various degrees of interference. And the parties involved often disagree on whether a given level of interference is harmless or will cause the victim receiver to malfunction. Reaching a decision on interference issues requires the FCC to make its way through a multi-dimensional maze of conflicting uncertainties. Here are some concrete issues that illuminate this all-too-common dynamic.
Fixed Ideas
Those ubiquitous sideways-facing dishes on towers and buildings are fixed‑microwave antennas. Equipment of this kind has operated reliably since the 1950s. The 6-GHz band, the lowest-frequency microwave band available today, is the only one capable of 100-kilometer hops, making it indispensable. Along with more pedestrian uses, the band carries safety-critical information: to coordinate trains, control pressure in oil and gas pipelines, balance the electric grid, manage water utilities, and route emergency telephone calls.
The red lines on this map of the 48 contiguous U.S. states show the location of existing 6-gigahertz fixed-microwave links, as recorded by Comsearch, which helps companies to avoid issues with radio interference. These links connect people in almost all areas, including far offshore in the Gulf of Mexico, where drilling platforms are common.
Comsearch
Four years ago, when the FCC proposed adding Wi-Fi to the 6-GHz band, all sides agreed that the vast majority of Wi-Fi devices would cause no trouble. Statistically, most would be outside the microwave antennas’ highly directional main beams, or on the wrong frequency, or shielded by buildings, terrain, and ground clutter.
The dispute centered on the small proportion of devices that might transmit on a frequency in use while being in the line-of-sight of a microwave antenna. The Wi-Fi proponents projected just under a billion devices, operating among 100,000 microwave receivers. The opponents pointed out that even a very small fraction of the many new transmitters could cause troubling numbers of interference events.
To mitigate the problem, the FCC adopted rules for an Automatic Frequency Control (AFC) system. A Wi-Fi device must either report its location to a central AFC database, which assigns it non-interfering frequencies for that location, or operate close to and under the control of an AFC-guided device. The AFC system will not be fully operational for another year or two, and disagreements persist about the details of its eventual operation.
More controversially, the FCC also authorized Wi-Fi devices without AFC, transmitting at will on any 6-GHz frequency from any geographic location—but only indoors and at no more than one-quarter of the maximum AFC-controlled power. The Wi-Fi proponents’ technical studies showed that attenuation from building walls would prevent interference. The microwave operators’ studies showed the opposite: that interference from uncontrolled indoor devices was virtually certain.
How could engineers, using the same equations, come to such different conclusions? These are a few of the ways in which their analyses differed:
Wi-Fi device power: A Wi-Fi device transmits in short bursts, active about 1/250th of the time, on average. The Wi-Fi proponents scaled down the power by a like amount, treating a device that transmits intermittently at, say, 250 milliwatts as though it transmitted continuously at 1 mW. The microwave operators argued that interference can occur only while the device is actually transmitting, so they calculated using the full power.
Building attenuation: A 6-GHz signal encounters substantial attenuation from concrete building walls and thermal windows, less from wood walls, and practically none from plain-glass windows. The Wi-Fi proponents took weighted averages over several building materials to calculate typical wall attenuations. The microwave operators reasoned that interference was most likely from an atypical Wi-Fi device behind plain glass, and they calculated accordingly, assuming a minimal amount of attenuation.
Path loss: In estimating the signal loss from a building that houses a Wi-Fi device to a microwave-receiving antenna, the Wi-Fi proponents used a standard propagation model that incorporates attenuation due to other buildings, ground clutter, and the like. The microwave operators were most concerned about a device located with open air between the building and the antenna, so they used free-space propagation in their calculations.
Using their preferred starting assumptions, the Wi-Fi proponents proved that Wi‑Fi devices over a wide range of typical situations present no risk of interference. Using a different set of assumptions, the microwave operators proved there is a large risk of interference from a small proportion of Wi-Fi devices in atypical locations, arguing that multiplying that small proportion by almost a billion Wi-Fi devices made interference virtually certain.
Up in the Air
Americans want their smartphones and tablets to have fast Internet access everywhere. That takes a lot of spectrum. Congress passed a statute in 2018 that told the FCC to find more—and specifically to consider 3.7 to 4.2 GHz, part of the C-band, used since the 1960s to receive satellite signals. The FCC partitioned the band in 2020, allocating 3.7 to 3.98 GHz for 5G mobile data. In early 2021, it auctioned the new 5G frequencies for US $81 billion, mostly to Verizon and AT&T. The auction winners were also expected to pay the satellite providers around $13 billion to compensate them for the costs of moving to other frequencies.
A nearby band at 4.2 to 4.4 GHz serves radar altimeters (also called radio altimeters), instruments that tell a pilot or an automatic landing system how high the aircraft is above the ground. The altimeter works by emitting downward radio waves that reflect off the ground and back up to a receiver in the device. The time for the round trip gives the altitude. Large planes operate two or three altimeters simultaneously, for redundancy.
Even though the altimeters use frequencies separated from the 5G band, they can still receive interference from 5G. That’s because every transmitter, including ones used for 5G, emits unwanted signals outside its assigned frequencies. Every receiver is likewise sensitive to signals outside its intended range, some more than others. Interference can occur if energy from a 5G transmitter falls within the sensitivity range of the receiver in an altimeter.
To make way for new 5G cellular services, the Federal Communications Commission reallocated part of the radio spectrum. That reallocation resulted in 5G transmissions that are close in frequency to a band used by aircraft radar altimeters.
The FCC regulates transmitter out-of-band emissions. In contrast, it has few rules on receiver out-of-band reception (although it recently opened a discussion on whether to expand them). Manufacturers generally design receivers to function reliably in their expected environments, which can leave them vulnerable if a new service appears in formerly quiet spectrum near the frequencies they receive on.
Aviation interests feared this outcome with the launch of C-band 5G, one citing the possibility of “ catastrophic impact with the ground, leading to multiple fatalities .” The FCC’s 5G order tersely dismissed concerns about altimeter interference, although it invited the aviation industry to study the matter further. The industry did so, renewing its concerns and requesting that the wireless carriers refrain from using 5G near airports. But this came after the wireless carriers had committed almost $100 billion and begun building out facilities.
Much as in the case of 6-GHz Wi-Fi, the 5G providers and aviation interests reached different predictions about interference by starting with different assumptions. Some key areas of disagreement were:
5G out-of-band emissions: The aviation interests assumed higher levels than the wireless carriers, which said the numbers in the aviation study levels exceeded FCC limits.
The FCC must regulate “in the public interest,” but the commissioners have to determine what that means in each case.
Off-channel sensitivity in altimeter receivers: There are several makes and models of altimeters in use, having varying receiver characteristics, leading to disagreements on which to include in the studies.
Altimeters in the same or other aircraft nearby. A busy airport has a lot of altimeters operating. Wireless carriers said these would overpower 5G interference. Aviation interests countered that multiple altimeters in the area would consume one another’s interference margin and leave them all more vulnerable to 5G.
Aircraft pitch and roll: Aviation interests argued that the changing angles of the aircraft as it approaches the runway can expose the altimeter receivers to more 5G signal.
Reflectivity of the ground: Aviation interests favored modeling with lower values of reflectivity, which reduce the received signal strength at the altimeter and hence increase its susceptibility to 5G interference.
The carriers temporarily paused 5G rollout near some airports, and the airlines canceled and rescheduled some flights. At this writing, the FAA is evaluating potentially affected aircraft, altimeters, and airport systems. Most likely, 5G will prevail. In the extremely improbable event that the FAA and the FCC were to agree that C-band 5G cannot operate safely near airports, the wireless carriers presumably would be entitled to a partial refund of their $81 billion auction payments.
These radio towers, which sit atop Black Mountain in Carmel Valley, Calif., include many drumlike antennas used for 6-gigahertz fixed-microwave links.
Shutterstock
Hard Decisions
Making complicated trade-offs has long been the job of the five FCC commissioners. They are political appointees, nominated by the president and confirmed by the Senate. The four now in office (there is a vacancy) are all lawyers. It has been decades since a commissioner had a technical background. The FCC has highly capable engineers on staff, but only in advisory roles. The commissioners have no obligation to take their advice.
Congress requires the FCC to regulate “in the public interest,” but the commissioners must determine what that means in each case. Legally, they can reach any result that has at least some support in the submissions, even if other submissions more strongly support an opposite result. Submissions to the FCC in both the 6-GHz and 5G matters conveyed sharp disagreement as to how much safety protection the public interest requires.
To fully protect 6-GHz microwave operations against interference from the small fraction of Wi-Fi devices in the line-of-sight of the microwave receivers would require degrading Wi-Fi service for large numbers of people. Similarly, eliminating any chance whatsoever of a catastrophic altimeter malfunction due to 5G interference might require turning off C-band 5G in some heavily populated areas.
The orders that authorized 6-GHz Wi-Fi and C-band 5G did not go that far and did not claim they had achieved zero risk. The order on 5G stated that altimeters had “all due protection.” In the 6-GHz case, with a federal appeals court deferring to its technical expertise, the FCC said it had “reduce[d] the possibility of harmful interference to the minimum that the public interest requires.”
These formulations make clear that safety is just one of several elements in the mix of public interests considered. Commissioners have to balance the goals of minimizing the risk of plane crashes and pipeline explosions against the demand for ubiquitous Internet access and Congress’s mandate to repurpose more spectrum.
In the end, the commissioners agreed with proponents’ claims that the risk of harmful interference from 6-GHz Wi-Fi is “insignificant,” although not zero, and similarly from 5G, not “likely…under…reasonably foreseeable scenarios”—conclusions that made it possible to offer the new services.
People like to think that the government puts the absolute safety of its citizens above all else. Regulation, though, like engineering, is an ever-shifting sequence of trade-offs. The officials who set highway speed limits know that lower numbers will save lives, but they also take into account motorists’ wishes to get to their destinations in a timely way. So it shouldn’t come as a great surprise that the FCC performs a similar balancing act.
This article appears in the July 2022 print issue as “Radio-Spectrum Turf Wars.”
From Your Site Articles