Mustaque Ahamad

Bing He, Mustaque Ahamad, Srijan Kumar

The spread of online misinformation threatens public health, democracy, and the broader society. While professional fact-checkers form the first line of defense by fact-checking popular false claims, they do not engage directly in conversations with misinformation spreaders. On the other hand, non-expert ordinary users act as eyes-on-the-ground who proactively counter misinformation -- recent research has shown that 96% counter-misinformation responses are made by ordinary users. However, research also found that 2/3 times, these responses are rude and lack evidence. This work seeks to create a counter-misinformation response generation model to empower users to effectively correct misinformation. This objective is challenging due to the absence of datasets containing ground-truth of ideal counter-misinformation responses, and the lack of models that can generate responses backed by communication theories. In this work, we create two novel datasets of misinformation and counter-misinformation response pairs from in-the-wild social media and crowdsourcing from college-educated students. We annotate the collected data to distinguish poor from ideal responses that are factual, polite, and refute misinformation. We propose MisinfoCorrect, a reinforcement learning-based framework that learns to generate counter-misinformation responses for an input misinformation post. The model rewards the generator to increase the politeness, factuality, and refutation attitude while retaining text fluency and relevancy. Quantitative and qualitative evaluation shows that our model outperforms several baselines by generating high-quality counter-responses. This work illustrates the promise of generative text models for social good -- here, to help create a safe and reliable information ecosystem. The code and data is accessible on https://github.com/claws-lab/MisinfoCorrect.

Bing He, Mustaque Ahamad, Srijan Kumar

Detecting bad actors is critical to ensure the safety and integrity of internet platforms. Several deep learning-based models have been developed to identify such users. These models should not only accurately detect bad actors, but also be robust against adversarial attacks that aim to evade detection. However, past deep learning-based detection models do not meet the robustness requirement because they are sensitive to even minor changes in the input sequence. To address this issue, we focus on (1) improving the model understanding capability and (2) enhancing the model knowledge such that the model can recognize potential input modifications when making predictions. To achieve these goals, we create a novel transformer-based classification model, called ROBAD (RObust adversary-aware local-global attended Bad Actor Detection model), which uses the sequence of user posts to generate user embedding to detect bad actors. Particularly, ROBAD first leverages the transformer encoder block to encode each post bidirectionally, thus building a post embedding to capture the local information at the post level. Next, it adopts the transformer decoder block to model the sequential pattern in the post embeddings by using the attention mechanism, which generates the sequence embedding to obtain the global information at the sequence level. Finally, to enrich the knowledge of the model, embeddings of modified sequences by mimicked attackers are fed into a contrastive-learning-enhanced classification layer for sequence prediction. In essence, by capturing the local and global information (i.e., the post and sequence information) and leveraging the mimicked behaviors of bad actors in training, ROBAD can be robust to adversarial attacks. Extensive experiments on Yelp and Wikipedia datasets show that ROBAD can effectively detect bad actors when under state-of-the-art adversarial attacks.

Govind Mittal, Sarthak Gupta, Shruti Wagle et al.

Radio remains a pervasive medium for mass information dissemination, with AM/FM stations reaching more Americans than either smartphone-based social networking or live television. Increasingly, radio broadcasts are also streamed online and accessed over the Internet. We present WavePulse, a framework that records, documents, and analyzes radio content in real-time. While our framework is generally applicable, we showcase the efficacy of WavePulse in a collaborative project with a team of political scientists focusing on the 2024 Presidential Elections. We use WavePulse to monitor livestreams of 396 news radio stations over a period of three months, processing close to 500,000 hours of audio streams. These streams were converted into time-stamped, diarized transcripts and analyzed to track answer key political science questions at both the national and state levels. Our analysis revealed how local issues interacted with national trends, providing insights into information flow. Our results demonstrate WavePulse's efficacy in capturing and analyzing content from radio livestreams sourced from the Web. Code and dataset can be accessed at \url{https://wave-pulse.io}.

Bing He, Mustaque Ahamad, Srijan Kumar

What should a malicious user write next to fool a detection model? Identifying malicious users is critical to ensure the safety and integrity of internet platforms. Several deep learning-based detection models have been created. However, malicious users can evade deep detection models by manipulating their behavior, rendering these models of little use. The vulnerability of such deep detection models against adversarial attacks is unknown. Here we create a novel adversarial attack model against deep user sequence embedding based classification models, which use the sequence of user posts to generate user embeddings and detect malicious users. In the attack, the adversary generates a new post to fool the classifier. We propose a novel end-to-end Personalized Text Generation Attack model, called PETGEN, that simultaneously reduces the efficacy of the detection model and generates posts that have several key desirable properties. Specifically, PETGEN generates posts that are personalized to the user's writing style, have knowledge about a given target context, are aware of the user's historical posts on the target context, and encapsulate the user's recent topical interests. We conduct extensive experiments on two real-world datasets (Yelp and Wikipedia, both with ground-truth of malicious users) to show that PETGEN significantly reduces the performance of popular deep user sequence embedding-based classification models. PETGEN outperforms five attack baselines in terms of text quality and attack efficacy in both white-box and black-box classifier settings. Overall, this work paves the path towards the next generation of adversary-aware sequence classification models.

Nicholas Micallef, Bing He, Srijan Kumar et al.

Fact checking by professionals is viewed as a vital defense in the fight against misinformation.While fact checking is important and its impact has been significant, fact checks could have limited visibility and may not reach the intended audience, such as those deeply embedded in polarized communities. Concerned citizens (i.e., the crowd), who are users of the platforms where misinformation appears, can play a crucial role in disseminating fact-checking information and in countering the spread of misinformation. To explore if this is the case, we conduct a data-driven study of misinformation on the Twitter platform, focusing on tweets related to the COVID-19 pandemic, analyzing the spread of misinformation, professional fact checks, and the crowd response to popular misleading claims about COVID-19. In this work, we curate a dataset of false claims and statements that seek to challenge or refute them. We train a classifier to create a novel dataset of 155,468 COVID-19-related tweets, containing 33,237 false claims and 33,413 refuting arguments.Our findings show that professional fact-checking tweets have limited volume and reach. In contrast, we observe that the surge in misinformation tweets results in a quick response and a corresponding increase in tweets that refute such misinformation. More importantly, we find contrasting differences in the way the crowd refutes tweets, some tweets appear to be opinions, while others contain concrete evidence, such as a link to a reputed source. Our work provides insights into how misinformation is organically countered in social platforms by some of their users and the role they play in amplifying professional fact checks.These insights could lead to development of tools and mechanisms that can empower concerned citizens in combating misinformation. The code and data can be found in http://claws.cc.gatech.edu/covid_counter_misinformation.html.

Zhengxian He, Mohit Narayan Rajput, Mustaque Ahamad

New security and privacy concerns arise due to the growing popularity of voice assistant (VA) deployments in home and enterprise networks. A number of past research results have demonstrated how malicious actors can use hidden commands to get VAs to perform certain operations even when a person may be in their vicinity. However, such work has not explored how compromised computers that are close to VAs can leverage the phone channel to exfiltrate data with the help of VAs. After characterizing the communication channel that is set up by commanding a VA to make a call to a phone number, we demonstrate how malware can encode data into audio and send it via the phone channel. Such an attack, which can be crafted remotely, at scale and at low cost, can be used to bypass network defenses that may be deployed against leakage of sensitive data. We use Dual-Tone Multi-Frequency tones to encode arbitrary binary data into audio that can be played over computer speakers and sent through a VA mediated phone channel to a remote system. We show that modest amounts of data can be transmitted with high accuracy with a short phone call lasting a few minutes. This can be done while making the audio nearly inaudible for most people by modulating it with a carrier with frequencies that are near the higher end of the human hearing range. Several factors influence the data transfer rate, including the distance between the computer and the VA, the ambient noise that may be present and the frequency of modulating carrier. With the help of a prototype built by us, we experimentally assess the impact of these factors on data transfer rates and transmission accuracy. Our results show that voice assistants in the vicinity of computers can pose new threats to data stored on such computers. These threats are not addressed by traditional host and network defenses. We briefly discuss possible mitigation ways.

Sharbani Pandit, Jienan Liu, Roberto Perdisci et al.

Mass robocalls affect millions of people on a daily basis. Unfortunately, most current defenses against robocalls rely on phone blocklists and are ineffective against caller ID spoofing. To enable the detection of spoofed robocalls, we propose a {\em virtual assistant} application that could be integrated on smartphones to automatically vet incoming calls. Similar to a human assistant, the virtual assistant can pick up an incoming call and screen it without user interruption to determine if the call is unwanted. Via a user study, we show that our virtual assistant is able to preserve the user experience of a typical phone call. At the same time, we show that our system can detect mass robocalls without negatively impacting legitimate callers.

Daniele Ucci, Roberto Perdisci, Jaewoo Lee et al.

Spam phone calls have been rapidly growing from nuisance to an increasingly effective scam delivery tool. To counter this increasingly successful attack vector, a number of commercial smartphone apps that promise to block spam phone calls have appeared on app stores, and are now used by hundreds of thousands or even millions of users. However, following a business model similar to some online social network services, these apps often collect call records or other potentially sensitive information from users' phones with little or no formal privacy guarantees. In this paper, we study whether it is possible to build a practical collaborative phone blacklisting system that makes use of local differential privacy (LDP) mechanisms to provide clear privacy guarantees. We analyze the challenges and trade-offs related to using LDP, evaluate our LDP-based system on real-world user-reported call records collected by the FTC, and show that it is possible to learn a phone blacklist using a reasonable overall privacy budget and at the same time preserve users' privacy while maintaining utility for the learned blacklist.

Bharat Srinivasan, Athanasios Kountouras, Najmeh Miramirkhani et al.

Technical Support Scams (TSS), which combine online abuse with social engineering over the phone channel, have persisted despite several law enforcement actions. The tactics used by these scammers have evolved over time and they have targeted an ever increasing number of technology brands. Although recent research has provided insights into TSS, these scams have now evolved to exploit ubiquitously used online services such as search and sponsored advertisements served in response to search queries. We use a data-driven approach to understand search-and-ad abuse by TSS to gain visibility into the online infrastructure that facilitates it. By carefully formulating tech support queries with multiple search engines, we collect data about both the support infrastructure and the websites to which TSS victims are directed when they search online for tech support resources. We augment this with a DNS-based amplification technique to further enhance visibility into this abuse infrastructure. By analyzing the collected data, we demonstrate that tech support scammers are (1) successful in getting major as well as custom search engines to return links to websites controlled by them, and (2) they are able to get ad networks to serve malicious advertisements that lead to scam pages. Our study period of 8 months uncovered over 9,000 TSS domains, of both passive and aggressive types, with minimal overlap between sets that are reached via organic search results and sponsored ads. Also, we found over 2,400 support domains which aid the TSS domains in manipulating organic search results. Moreover, we found little overlap with domains that are reached via abuse of domain parking and URL-shortening services which was investigated previously. Thus, investigation of search-and-ad abuse provides new insights into TSS tactics and helps detect previously unknown abuse infrastructure that facilitates these scams.

Srishti Gupta, Payas Gupta, Mustaque Ahamad et al.

With the convergence of Internet and telephony, new applications (e.g., WhatsApp) have emerged as an important means of communication for billions of users. These applications are becoming an attractive medium for attackers to deliver spam and carry out more targeted attacks. Since such applications rely on phone numbers, we explore the feasibility, automation, and scalability of phishing attacks that can be carried out by abusing a phone number. We demonstrate a novel system that takes a potential victim's phone number as an input, leverages information from applications like Truecaller and Facebook about the victim and his / her social network, checks the presence of phone number's owner (victim) on the attack channels (over-the-top or OTT messaging applications, voice, e-mail, or SMS), and finally targets the victim on the chosen channel. As a proof of concept, we enumerate through a random pool of 1.16 million phone numbers. By using information provided by popular applications, we show that social and spear phishing attacks can be launched against 51,409 and 180,000 users respectively. Furthermore, voice phishing or vishing attacks can be launched against 722,696 users. We also found 91,487 highly attractive targets who can be attacked by crafting whaling attacks. We show the effectiveness of one of these attacks, phishing, by conducting an online roleplay user study. We found that social (69.2%) and spear (54.3%) phishing attacks are more successful than non-targeted phishing attacks (35.5%) on OTT messaging applications. Although similar results were found for other mediums like e-mail, we demonstrate that due to the significantly increased user engagement via new communication applications and the ease with which phone numbers allow collection of information necessary for these attacks, there is a clear need for better protection of OTT messaging applications.