Post Session Report on Universal Acceptance and Multilingual Internet at BIT University under CyberPeace Center of Excellence (CCoE)

Introduction

The fast-paced development of technology and the wider use of social media platforms have led to the rapid dissemination of misinformation with characteristics such as diffusion, fast propagation speed, wide influence, and deep impact through these platforms. Social Media Algorithms and their decisions are often perceived as a black box introduction that makes it impossible for users to understand and recognise how the decision-making process works. 

Social media algorithms may unintentionally promote false narratives that garner more interactions, further reinforcing the misinformation cycle and making it harder to control its spread within vast, interconnected networks. Algorithms judge the content based on the metrics, which is user engagement. It is the prerequisite for algorithms to serve you the best. Hence, algorithms or search engines enlist relevant items you are more likely to enjoy. This process, initially, was created to cut the clutter and provide you with the best information. However, sometimes it results in unknowingly widespread misinformation due to the viral nature of information and user interactions. 

Analysing the Algorithmic Architecture of Misinformation

Social media algorithms, designed to maximize user engagement, can inadvertently promote misinformation due to their tendency to trigger strong emotions, creating echo chambers and filter bubbles. These algorithms prioritize content based on user behaviour, leading to the promotion of emotionally charged misinformation. Additionally, the algorithms prioritize content that has the potential to go viral, which can lead to the spread of false or misleading content faster than corrections or factual content.

Additionally, popular content is amplified by platforms, which spreads it faster by presenting it to more users. Limited fact-checking efforts are particularly difficult since, by the time they are reported or corrected, erroneous claims may have gained widespread acceptance due to delayed responses. Social media algorithms find it difficult to distinguish between real people and organized networks of troll farms or bots that propagate false information. This creates a vicious loop where users are constantly exposed to inaccurate or misleading material, which strengthens their convictions and disseminates erroneous information through networks.

Though algorithms, primarily, aim to enhance user engagement by curating content that aligns with the user's previous behaviour and preferences. Sometimes this process leads to "echo chambers," where individuals are exposed mainly to information that reaffirms their beliefs which existed prior, effectively silencing dissenting voices and opposing viewpoints. This curated experience reduces exposure to diverse opinions and amplifies biased and polarising content, making it arduous for users to discern credible information from misinformation. Algorithms feed into a feedback loop that continuously gathers data from users' activities across digital platforms, including websites, social media, and apps. This data is analysed to optimise user experiences, making platforms more attractive. While this process drives innovation and improves user satisfaction from a business standpoint, it also poses a danger in the context of misinformation. The repetitive reinforcement of user preferences leads to the entrenchment of false beliefs, as users are less likely to encounter fact-checks or corrective information.

Moreover, social networks and their sheer size and complexity today exacerbate the issue. With billions of users participating in online spaces, misinformation spreads rapidly, and attempting to contain it—such as by inspecting messages or URLs for false information—can be computationally challenging and inefficient. The extensive amount of content that is shared daily means that misinformation can be propagated far quicker than it can get fact-checked or debunked.

Understanding how algorithms influence user behaviour is important to tackling misinformation. The personalisation of content, feedback loops, the complexity of network structures, and the role of superspreaders all work together to create a challenging environment where misinformation thrives. Hence, highlighting the importance of countering misinformation through robust measures. 

The Role of Regulations in Curbing Algorithmic Misinformation

The EU's Digital Services Act (DSA) applicable in the EU is one of the regulations that aims to increase the responsibilities of tech companies and ensure that their algorithms do not promote harmful content. These regulatory frameworks play an important role they can be used to establish mechanisms for users to appeal against the algorithmic decisions and ensure that these systems do not disproportionately suppress legitimate voices. Independent oversight and periodic audits can ensure that algorithms are not biased or used maliciously. Self-regulation and Platform regulation are the first steps that can be taken to regulate misinformation. By fostering a more transparent and accountable ecosystem, regulations help mitigate the negative effects of algorithmic misinformation, thereby protecting the integrity of information that is shared online. In the Indian context, the Intermediary Guidelines, 2023, Rule 3(1)(b)(v) explicitly prohibits the dissemination of misinformation on digital platforms. The ‘Intermediaries’ are obliged to ensure reasonable efforts to prevent users from hosting, displaying, uploading, modifying, publishing, transmitting, storing, updating, or sharing any information related to the 11 listed user harms or prohibited content. This rule aims to ensure platforms identify and swiftly remove misinformation, and false or misleading content.

Cyberpeace Outlook

Understanding how algorithms prioritise content will enable users to critically evaluate the information they encounter and recognise potential biases. Such cognitive defenses can empower individuals to question the sources of the information and report misleading content effectively. In the future of algorithms in information moderation, platforms should evolve toward more transparent, user-driven systems where algorithms are optimised not just for engagement but for accuracy and fairness. Incorporating advanced AI moderation tools, coupled with human oversight can improve the detection and reduction of harmful and misleading content. Collaboration between regulatory bodies, tech companies, and users will help shape the algorithms landscape to promote a healthier, more informed digital environment.

References:

• https://www.advancedsciencenews.com/misformation-spreads-like-a-nuclear-reaction-on-the-internet/
• https://www.niemanlab.org/2024/09/want-to-fight-misinformation-teach-people-how-algorithms-work/ 
• Press Release: Press Information Bureau (pib.gov.in) 

‍

Executive Summary:

A viral image circulating on social media claims to show a Hindu Sadhvi marrying a Muslim man; however, this claim is false. A thorough investigation by the Cyberpeace Research team found that the image has been digitally manipulated. The original photo, which was posted by Balmukund Acharya, a BJP MLA from Jaipur, on his official Facebook account in December 2023, he was posing with a Muslim man in his election office. The man wearing the Muslim skullcap is featured in several other photos on Acharya's Instagram account, where he expressed gratitude for the support from the Muslim community. Thus, the claimed image of a marriage between a Hindu Sadhvi and a Muslim man is digitally altered.

Claims:

An image circulating on social media claims to show a Hindu Sadhvi marrying a Muslim man. 

‍

Fact Check:

Upon receiving the posts, we reverse searched the image to find any credible sources. We found a photo posted by Balmukund Acharya Hathoj Dham on his facebook page on 6 December 2023. 

This photo is digitally altered and posted on social media to mislead. We also found several different photos with the skullcap man where he was featured.

‍

We also checked for any AI fabrication in the viral image. We checked using a detection tool named, “content@scale” AI Image detection.  This tool found the image to be 95% AI Manipulated.

‍

We also checked with another detection tool for further validation named, “isitai” image detection tool. It found the image to be 38.50% of AI content, which concludes to the fact that the image is manipulated and doesn’t support the claim made. Hence, the viral image is fake and misleading.

Conclusion:

The lack of credible source and the detection of AI manipulation in the image explains that the viral image claiming to show a Hindu Sadhvi marrying a Muslim man is false. It has been digitally altered. The original image features BJP MLA Balmukund Acharya posing with a Muslim man, and there is no evidence of the claimed marriage.

• Claim: An image circulating on social media claims to show a Hindu Sadhvi marrying a Muslim man. 
• Claimed on: X (Formerly known as Twitter)
• Fact Check: Fake & Misleading

Executive Summary:

In late 2024 an Indian healthcare provider experienced a severe cybersecurity attack that demonstrated how powerful AI ransomware is. This blog discusses the background to the attack, how it took place and the effects it caused (both medical and financial), how organisations reacted, and the final result of it all, stressing on possible dangers in the healthcare industry with a lack of sufficiently adequate cybersecurity measures in place. The incident also interrupted the normal functioning of business and explained the possible economic and image losses from cyber threats. Other technical results of the study also provide more evidence and analysis of the advanced AI malware and best practices for defending against them. 

1. Introduction

The integration of artificial intelligence (AI) in cybersecurity has revolutionised both defence mechanisms and the strategies employed by cybercriminals. AI-powered attacks, particularly ransomware, have become increasingly sophisticated, posing significant threats to various sectors, including healthcare. This report delves into a case study of an AI-powered ransomware attack on a prominent Indian healthcare provider in 2024, analysing the attack's execution, impact, and the subsequent response, along with key technical findings.

2. Background 

 In late 2024, a leading healthcare organisation in India which is involved in the research and development of AI techniques fell prey to a ransomware attack that was AI driven to get the most out of it. With many businesses today relying on data especially in the healthcare industry that requires real-time operations, health care has become the favourite of cyber criminals. AI aided attackers were able to cause far more detailed and damaging attack that severely affected the operation of the provider whilst jeopardising the safety of the patient information.  

 3. Attack Execution 

The attack began with the launch of a phishing email designed to target a hospital administrator. They received an email with an infected attachment which when clicked in some cases injected the AI enabled ransomware into the hospitals network. AI incorporated ransomware was not as blasé as traditional ransomware, which sends copies to anyone, this studied the hospital’s IT network. First, it focused and targeted important systems which involved implementation of encryption such as the electronic health records and the billing departments. 

The fact that the malware had an AI feature allowed it to learn and adjust its way of propagation in the network, and prioritise the encryption of most valuable data. This accuracy did not only increase the possibility of the potential ransom demand but also it allowed reducing the risks of the possibility of early discovery. 

 4. Impact 

•  The consequences of the attack were immediate and severe: The consequences of the attack were immediate and severe.
•  Operational Disruption: The centralization of important systems made the hospital cease its functionality through the acts of encrypting the respective components. Operations such as surgeries, routine medical procedures and admitting of patients were slowed or in some cases referred to other hospitals. 
•  Data Security: Electronic patient records and associated billing data became off-limit because of the vulnerability of patient confidentiality. The danger of data loss was on the verge of becoming permanent, much to the concern of both the healthcare provider and its patients. 
•  Financial Loss: The attackers asked for 100 crore Indian rupees (approximately 12 USD million) for the decryption key. Despite the hospital not paying for it, there were certain losses that include the operational loss due to the server being down, loss incurred by the patients who were affected in one way or the other, loss incurred in responding to such an incident and the loss due to bad reputation. 

5. Response

As soon as the hotel’s management was informed about the presence of ransomware, its IT department joined forces with cybersecurity professionals and local police. The team decided not to pay the ransom and instead recover the systems from backup. Despite the fact that this was an ethically and strategically correct decision, it was not without some challenges. Reconstruction was gradual, and certain elements of the patients’ records were permanently erased. 

 In order to avoid such attacks in the future, the healthcare provider put into force several organisational and technical actions such as network isolation and increase of cybersecurity measures. Even so, the attack revealed serious breaches in the provider’s IT systems security measures and protocols. 

6. Outcome

The attack had far-reaching consequences:

• Financial Impact: A healthcare provider suffers a lot of crashes in its reckoning due to substantial service disruption as well as bolstering cybersecurity and compensating patients. 
•  Reputational Damage: The leakage of the data had a potential of causing a complete loss of confidence from patients and the public this affecting the reputation of the provider. This, of course, had an effect on patient care, and ultimately resulted in long-term effects on revenue as patients were retained. 
•  Industry Awareness: The breakthrough fed discussions across the country on how to improve cybersecurity provisions in the healthcare industry. It woke up the other care providers to review and improve their cyber defence status. 

7. Technical Findings

The AI-powered ransomware attack on the healthcare provider revealed several technical vulnerabilities and provided insights into the sophisticated mechanisms employed by the attackers. These findings highlight the evolving threat landscape and the importance of advanced cybersecurity measures.

7.1 Phishing Vector and Initial Penetration

• Sophisticated Phishing Tactics: The phishing email was crafted with precision, utilising AI to mimic the communication style of trusted contacts within the organisation. The email bypassed standard email filters, indicating a high level of customization and adaptation, likely due to AI-driven analysis of previous successful phishing attempts.
• Exploitation of Human Error: The phishing email targeted an administrative user with access to critical systems, exploiting the lack of stringent access controls and user awareness. The successful penetration into the network highlighted the need for multi-factor authentication (MFA) and continuous training on identifying phishing attempts.

7.2 AI-Driven Malware Behavior

• Dynamic Network Mapping: Once inside the network, the AI-powered malware executed a sophisticated mapping of the hospital's IT infrastructure. Using machine learning algorithms, the malware identified the most critical systems—such as Electronic Health Records (EHR) and the billing system—prioritising them for encryption. This dynamic mapping capability allowed the malware to maximise damage while minimising its footprint, delaying detection.
• Adaptive Encryption Techniques: The malware employed adaptive encryption techniques, adjusting its encryption strategy based on the system's response. For instance, if it detected attempts to isolate the network or initiate backup protocols, it accelerated the encryption process or targeted backup systems directly, demonstrating an ability to anticipate and counteract defensive measures.
• Evasive Tactics: The ransomware utilised advanced evasion tactics, such as polymorphic code and anti-forensic features, to avoid detection by traditional antivirus software and security monitoring tools. The AI component allowed the malware to alter its code and behaviour in real time, making signature-based detection methods ineffective.

7.3 Vulnerability Exploitation

• Weaknesses in Network Segmentation: The hospital’s network was insufficiently segmented, allowing the ransomware to spread rapidly across various departments. The malware exploited this lack of segmentation to access critical systems that should have been isolated from each other, indicating the need for stronger network architecture and micro-segmentation.
• Inadequate Patch Management: The attackers exploited unpatched vulnerabilities in the hospital’s IT infrastructure, particularly within outdated software used for managing patient records and billing. The failure to apply timely patches allowed the ransomware to penetrate and escalate privileges within the network, underlining the importance of rigorous patch management policies.

7.4 Data Recovery and Backup Failures

• Inaccessible Backups: The malware specifically targeted backup servers, encrypting them alongside primary systems. This revealed weaknesses in the backup strategy, including the lack of offline or immutable backups that could have been used for recovery. The healthcare provider’s reliance on connected backups left them vulnerable to such targeted attacks.
• Slow Recovery Process: The restoration of systems from backups was hindered by the sheer volume of encrypted data and the complexity of the hospital’s IT environment. The investigation found that the backups were not regularly tested for integrity and completeness, resulting in partial data loss and extended downtime during recovery.

7.5 Incident Response and Containment

• Delayed Detection and Response: The initial response was delayed due to the sophisticated nature of the attack, with traditional security measures failing to identify the ransomware until significant damage had occurred. The AI-powered malware’s ability to adapt and camouflage its activities contributed to this delay, highlighting the need for AI-enhanced detection and response tools.
• Forensic Analysis Challenges: The anti-forensic capabilities of the malware, including log wiping and data obfuscation, complicated the post-incident forensic analysis. Investigators had to rely on advanced techniques, such as memory forensics and machine learning-based anomaly detection, to trace the malware’s activities and identify the attack vector.

8. Recommendations Based on Technical Findings

To prevent similar incidents, the following measures are recommended:

1. AI-Powered Threat Detection: Implement AI-driven threat detection systems capable of identifying and responding to AI-powered attacks in real time. These systems should include behavioural analysis, anomaly detection, and machine learning models trained on diverse datasets.
2. Enhanced Backup Strategies: Develop a more resilient backup strategy that includes offline, air-gapped, or immutable backups. Regularly test backup systems to ensure they can be restored quickly and effectively in the event of a ransomware attack.
3. Strengthened Network Segmentation: Re-architect the network with robust segmentation and micro-segmentation to limit the spread of malware. Critical systems should be isolated, and access should be tightly controlled and monitored.
4. Regular Vulnerability Assessments: Conduct frequent vulnerability assessments and patch management audits to ensure all systems are up to date. Implement automated patch management tools where possible to reduce the window of exposure to known vulnerabilities.
5. Advanced Phishing Defences: Deploy AI-powered anti-phishing tools that can detect and block sophisticated phishing attempts. Train staff regularly on the latest phishing tactics, including how to recognize AI-generated phishing emails.

9. Conclusion

The AI empowered ransomware attack on the Indian healthcare provider in 2024 makes it clear that the threat of advanced cyber attacks has grown in the healthcare facilities. Sophisticated technical brief outlines the steps used by hackers hence underlining the importance of ongoing active and strong security. This event is a stark message to all about the importance of not only remaining alert and implementing strong investments in cybersecurity but also embarking on the formulation of measures on how best to counter such incidents with limited harm. AI is now being used by cybercriminals to increase the effectiveness of the attacks they make and it is now high time all healthcare organisations ensure that their crucial systems and data are well protected from such attacks.

‍