These dangerous scammers don’t even bother to hide their crimes

One hundred dollar bill Benjamin Franklin portrait looks behind brown craft ripped paper

Most scammers and cybercriminals operate in the digital shadows and don’t want you to know how they make money. But that’s not the case for the Yahoo Boys, a loose collective of young men in West Africa who are some of the web’s most prolific—and increasingly dangerous—scammers.

Thousands of people are members of dozens of Yahoo Boy groups operating across Facebook, WhatsApp, and Telegram, a WIRED analysis has found. The scammers, who deal in types of fraud that total hundreds of millions of dollars each year, also have dozens of accounts on TikTok, YouTube, and the document-sharing service Scribd that are getting thousands of views.

Inside the groups, there’s a hive of fraudulent activity with the cybercriminals often showing their faces and sharing ways to scam people with other members. They openly distribute scripts detailing how to blackmail people and how to run sextortion scams—that have driven people to take their own lives—sell albums with hundreds of photographs, and advertise fake social media accounts. Among the scams, they’re also using AI to create fake “nude” images of people and real-time deepfake video calls.

The Yahoo Boys don’t disguise their activity. Many groups use “Yahoo Boys” in their name as well as other related terms. WIRED’s analysis found 16 Yahoo Boys Facebook groups with almost 200,000 total members, a dozen WhatsApp channels, around 10 Telegram channels, 20 TikTok accounts, a dozen YouTube accounts, and more than 80 scripts on Scribd. And that’s just the tip of the iceberg.

Broadly, the companies do not allow content on their platforms that encourages or promotes criminal behavior. The majority of the Yahoo Boys accounts and groups WIRED identified were removed after we contacted the companies about the groups’ overt existence. Despite these removals, dozens more Yahoo Boys groups and accounts remain online.

“They’re not hiding under different names,” says Kathy Waters, the co-founder and executive director of the nonprofit Advocating Against Romance Scammers, which has tracked the Yahoo Boys for years. Waters says the social media companies are essentially providing the Yahoo Boys with “free office space” to organize and conduct their activities. “They’re selling scripts, selling photos, identifications of people, all online, all on the social media platforms,” she says. “Why these accounts still remain is beyond me.”

The Yahoo Boys aren’t a single, organized group. Instead, they’re a collection of thousands of scammers who work individually or in clusters. Often based in Nigeria, their name comes from formerly targeting users of Yahoo services, with links back to the Nigerian Prince email scams of old. Groups in West Africa can be often organized in various confraternities, which are cultish gangs.

“Yahoo is a set of knowledge that allows you to conduct scams,” says Gary Warner, the director of intelligence at DarkTower and director of the University of Alabama at Birmingham’s Computer Forensics Research Laboratory. While there are different levels of sophistication of Yahoo Boys, Warner says, many simply operate from their phones. “Most of these threat actors are only using one device,” he says.

The Yahoo Boys run dozens of scams—from romance fraud to business email compromise. When making contact with potential victims, they’ll often “bomb” people by sending hundreds of messages to dating app accounts or Facebook profiles. “They will say anything they can in order to get the next dime in their pocket,” Waters says.

Searching for the Yahoo Boys on Facebook brings up two warnings: Both say the results may be linked to fraudulent activity, which isn’t allowed on the website. Clicking through the warnings reveals Yahoo Boy groups with thousands of members—one had more than 70,000.

Within the groups—alongside posts selling SIM cards and albums with hundreds of pictures—many of the scammers push people toward other messaging platforms such as Meta’s WhatsApp or Telegram. Here, the Yahoo Boys are at their most bold. Some groups and channels on the two platforms receive hundreds of posts per day and are part of their wider web of operations.

After WIRED asked Facebook about the 16 groups we identified, the company removed them, and some WhatsApp groups were deactivated. “Scammers use every platform available to them to defraud people and constantly adapt to avoid getting caught,” says Al Tolan, a Meta spokesperson. They did not directly address the accounts that were removed or that they were easy to find. “Purposefully exploiting others for money is against our policies, and we take action when we become aware of it,” Tolan says. “We continue to invest in technology and cooperate with law enforcement so they can prosecute scammers. We also actively share tips on how people can protect themselves, their accounts, and avoid scams.”

Groups on Telegram were removed after WIRED messaged the company’s press office; however, the platform did not respond about why it had removed them.

Across all types of social media, Yahoo Boys scammers share “scripts” that they use to socially manipulate people—these can run to thousands of words long and can be copied and pasted to different victims. Many have been online for years. “I’ve seen some scripts that are 30 and 60 layers deep, before the scammer actually would have to go and think of something else to say,” says Ronnie Tokazowski, the chief fraud fighter at Intelligence for Good, which works with cybercrime victims. “It’s 100 percent how they’ll manipulate the people,” Tokazowski says.

Among the many scams, they pretend to be military officers, people offering “hookups,” the FBI, doctors, and people looking for love. One “good morning” script includes around a dozen messages the scammers can send to their targets. “In a world full of deceit and lies, I feel lucky when see the love in your eyes. Good morning,” one says. But things get much darker.

https://arstechnica.com/?p=2022009




Microsoft plans to lock down Windows DNS like never before. Here’s how.

Microsoft plans to lock down Windows DNS like never before. Here’s how.
Getty Images

Translating human-readable domain names into numerical IP addresses has long been fraught with gaping security risks. After all, lookups are rarely end-to-end encrypted. The servers providing domain name lookups provide translations for virtually any IP address—even when they’re known to be malicious. And many end-user devices can easily be configured to stop using authorized lookup servers and instead use malicious ones.

Microsoft on Friday provided a peek at a comprehensive framework that aims to sort out the Domain Name System (DNS) mess so that it’s better locked down inside Windows networks. It’s called ZTDNS (zero trust DNS). Its two main features are (1) encrypted and cryptographically authenticated connections between end-user clients and DNS servers and (2) the ability for administrators to tightly restrict the domains these servers will resolve.

Clearing the minefield

One of the reasons DNS has been such a security minefield is that these two features can be mutually exclusive. Adding cryptographic authentication and encryption to DNS often obscures the visibility admins need to prevent user devices from connecting to malicious domains or detect anomalous behavior inside a network. As a result, DNS traffic is either sent in clear text or it’s encrypted in a way that allows admins to decrypt it in transit through what is essentially an adversary-in-the-middle attack.

Admins are left to choose between equally unappealing options: (1) route DNS traffic in clear text with no means for the server and client device to authenticate each other so malicious domains can be blocked and network monitoring is possible, or (2) encrypt and authenticate DNS traffic and do away with the domain control and network visibility.

ZTDNS aims to solve this decades-old problem by integrating the Windows DNS engine with the Windows Filtering Platform—the core component of the Windows Firewall—directly into client devices.

Jake Williams, VP of research and development at consultancy Hunter Strategies, said the union of these previously disparate engines would allow updates to be made to the Windows firewall on a per-domain name basis. The result, he said, is a mechanism that allows organizations to, in essence, tell clients “only use our DNS server, that uses TLS, and will only resolve certain domains.” Microsoft calls this DNS server or servers the “protective DNS server.”

By default, the firewall will deny resolutions to all domains except those enumerated in allow lists. A separate allow list will contain IP address subnets that clients need to run authorized software. Key to making this work at scale inside an organization with rapidly changing needs. Networking security expert Royce Williams (no relation to Jake Williams) called this a “sort of a bidirectional API for the firewall layer, so you can both trigger firewall actions (by input *to* the firewall), and trigger external actions based on firewall state (output *from* the firewall). So instead of having to reinvent the firewall wheel if you are an AV vendor or whatever, you just hook into WFP.”

https://arstechnica.com/?p=2021987




Verizon 2024 Data Breach Report shows the risk of the human element

Verizon Business released its 17th-annual Data Breach Investigations Report (DBIR), highlighting the role that the human element plays in cyber threats. This report examined 30,458 security incidents as well as 10,626 verified breaches in 2023, representing a two-fold increase from 2022. Out of the breaches analyzed, more than two-thirds (68%) included a non-malicious human element — in other words, these incidents involved insider errors or people falling for social engineering schemes.

This percentage remains consistent with last year’s, suggesting that the human element remains a steady risk concern. However, reporting practices improved, as 20% of individuals recognized and reported phishing in simulated exercises, and 11% of individuals who clicked a malicious email reported it. 

Another notable finding from the report was the increase in vulnerability exploitation. Exploiting vulnerabilities as an initial entry point accounted for 14% of all breaches, representing a volume three times (180%) greater than 2023. According to the report, this increase was driven by zero-day vulnerabilities that ransomware actors leveraged. 

Other key findings include: 

  • 32% of breaches included a form of extortion, including ransomware.
  • Between 24% and 25% of financially motivated security events involved pretexting over the past two years. 
  • Over the last decade, 31% of breaches involved the use of stolen credentials.

Security leaders weigh in 

Saeed Abbasi, Manager, Vulnerability Research at Qualys Threat Research Unit:

“The 2024 Verizon Data Breach Investigations Report (DBIR) highlights cyber threats that are evolving and increasingly complex in our interconnected world. These findings spotlight a crucial theme: today’s cyber threats are’ dynamic and increasingly sophisticated in nature. 

“Here are my takeaways this year: 

  • Adaptive threat landscape: The report details a notable increase in ransomware, extortion techniques, and vulnerability exploitation, showing that cybercriminals are becoming more adaptive and opportunistic. They effectively utilize everything from zero-day vulnerabilities to social engineering tactics like phishing to penetrate systems. 
  • Convergence of threats: It also notes an evolution of ransomware into more complex forms of extortion, marking a convergence of threats where different attack methods merge into hybrid tactics. This convergence complicates organizations’ ability to predict and defend against attacks as the distinctions between attack types become increasingly blurred. 
  • Human element in cybersecurity: This highlights the rapid rate at which individuals fall for phishing scams, the DBIR underscores the critical importance of human behavior in cybersecurity. It advocates for a dual approach that focuses on technological defenses and emphasizes the need for comprehensive user education and behavioral adjustments to bolster security. 
  • Strategic vulnerability management and holistic defense mechanisms: The 2024 Verizon DBIR emphasizes a critical increase in vulnerability exploitations, highlighting the need for urgent, strategic vulnerability management. We advise organizations to implement comprehensive, proactive strategies, including agent-based and agent-less security measures, to preempt potential breaches. Additionally, organizations require a multi-layered defense strategy, integrating advanced detection tools, zero-trust frameworks, and rapid patch management. 

“Given the increasing complexity and interconnectedness of supply chains, this holistic approach to cybersecurity is essential. These networks are often targeted by cyber threats, affecting not just individual organizations but also extending to third-party interactions and the broader supply chain. 

  • AI, Machine Learning, and Quantum Computing Threats: The proliferation of AI and machine learning is expected to be leveraged by both defenders and attackers. AI can swiftly predict and counteract attacks but may also be used to develop more sophisticated cyber threats. Additionally, as quantum computing advances, it poses a potential risk to current cryptographic protocols. Organizations should prepare for this by developing quantum-resistant cryptography to safeguard data against future threats.” 

Patrick Harr, CEO at SlashNext:

“With the rapid growth of AI technology, combined with limited regulation, it’s important for the tech industry to develop tools and processes that can assist in protecting AI technology systems.

“Everything in security needs to become more human ID-centric rather than network-centric. At the end of the day, we are far better off by providing access through human identity-centric methods and using AI to make that human a super-human. So rather than relying on a training simulation approach for users, we can rely on AI augmentation for that, so users don’t have to be tricked into clicking on bad phishing links, for example. 

“We have to shift our posture from a network-centric to a human-centric security posture. We will put an AI bubble around the user to become a super-human with an extra pair of computer vision eyes, and an ability to listen with spoken language contextualization by using AI. Everyone has talked about a personal co-pilot to help from a security posture, and we will see the rise of these AI co-pilots to augment humans and help users make the best decisions.  

“This problem will not go away and will only get worse. Anywhere there is money and opportunity and data, which is across every industry, there will be attacks. This is a horizontal problem for all industries, not a vertical problem. The bad guys will always look for wherever the most sensitive data is based to target their attacks.” 

Agnidipta Sarkar, Vice President CISO Advisory at ColorTokens:

“With the increase in digital business-as-usual, cybersecurity practitioners are already feeling lost in a deluge of inaccurate information from mushrooming multiple cybersecurity solutions coupled with a lack of cybersecurity architecture and design practices, resulting in porous cyber defenses. Business leaders are realizing that investments in microsegmentation will force the IT and security teams to begin developing digital business context-based cybersecurity architecture and design because microsegmentation is the last line of defense during a cyber-attack. Security and risk leaders will leverage the pan-optic visualization capability of microsegmentation to build immediate cyber defenses to protect digital business as usual, even during severe cyber-attacks.” 

Nick Rago, Vice President, Product Strategy at Salt Security:

“As architectures become increasingly complex, combined with more dependencies on third party code and services, supply chain attacks targeting software dependencies and operational third-party providers will continue to escalate. Especially as threats actor techniques become more stealth and harder to detect.

“In the wake of successful attacks over the past year, social engineering attacks are set to continue and with more sophistication. As a result, zero-trust mindsets should be applied to every communication medium, including corporate email, text message, or phone call. The education and re-education of employees must continue. 

“API attacks will also continue to increase at an alarming rate as organizations struggle to manage the chaos of API sprawl stemming from API-first innovation and digitalization. On the flip side, it is likely organizations will allocate more budget towards API security in the new year given its increased importance. In Salt Security’s State of the CISO report, 95% of CISOs surveyed said API security is a planned priority over the next 24 months.”

Dana Simberkoff, Chief Risk, Privacy and Information Security Officer at AvePoint:

“AI is coming and resistance is futile. While we see the great potential AI can have to help us in our work, we must make sure that we take advantage of these technologies responsibly and securely. In light of this, Security and Privacy professionals must work with their IT and business counterparts to develop and implement Generative AI Acceptable use policies. This should include data privacy and confidentiality, access to generative AI, and responsible use of generative AI. Putting these guardrails in place is critical. 

“In addition to developing acceptable use policies, ensure that you have ongoing training for employees so that they are aware and can act responsibly. Especially given how quickly applications of AI and machine learning have impacted our work, and how quickly this technology changes, security and privacy teams need to be agile in the new year.

“Successful adoption of AI in a security and privacy centric way will be as good as the basic data governance and life cycle management program you’ve implemented in your organization. As we say and have said for many years with regards to migration to the cloud: if you put garbage in, you’ll get garbage out. So, it’s important to clean up your data and make sure its properly governed before serving it up to AI on a silver platter. Otherwise, you may end up finding that security by obscurity is no longer a fallback defense.” 

https://www.securitymagazine.com/articles/100629-verizon-2024-data-breach-report-shows-the-risk-of-the-human-element




Zscaler elimina la necessità della segmentazione basata su firewall


Zscaler ha annunciato l’acquisizione di Airgap Networks, azienda che fornisce segmentazione agentless per ambienti IT e OT. La nuova tecnologia cambierà il modo in cui le aziende implementano la segmentazione Zero Trust nelle infrastrutture critiche, eliminando la necessità di firewall.

“Da sempre la segmentazione è stata realizzata con tecnologie di rete centrate sull’IP e ormai obsolete, come il NAC e i firewall est-ovest, gestite attraverso costrutti complessi come le ACL basate su indirizzi MAC, IP e VLAN. Questa complessità mette a dura prova i team operativi di rete, costretti a scrivere, mantenere e aggiornare innumerevoli ACL o regole di firewall interni, affrontando al contempo le inevitabili configurazioni errate che interrompono le applicazioni critiche per l’azienda o lasciano lacune nella copertura della segmentazione” ha affermato Naresh Kumar, VP Product Management di Zscaler.

La complessità dei firewall est-ovest fa sì che la maggior parte dei progetti di segmentazione di rete non siano mai del tutto implementati e, in generale, le policy di segmentazione tendono a fallire a causa dello spostamento dei carichi di lavoro e dei cambiamenti degli ambienti.

rete web network

L’approccio innovativo di Airgap permette di avere visibilità completa e implementare policy stringenti su ogni endpoint connesso senza la necessità di aggiungere software, riducendo la complessità generale di gestione.

Airgap Networks usa un’architettura proxy intelligente del Dynamic Host Configuration Protocol (DHCP) per isolare i dispositivi e controllare dinamicamente l’accesso in base all’identità e al contesto, riducendo i rischi. La soluzione ripensa totalmente la complessità degli approcci legacy di segmentazione della rete fornendo una segmentazione più robusta e con maggiore semplicità operativa.

La tecnologia agentless applica i principi Zero trust al traffico laterale dei dispositivi riducendo la superficie di attacco interna e il rischio di spostamento laterale delle minacce nelle reti OT e nei campus. La soluzione consente inoltre di individuare i dispositivi in tempo reale e di fare enforcement in linea. Agendo come un kill switch per il ransomware, la tecnologia agentless disabilita le comunicazioni non essenziali per impedire i movimenti laterali.

Grazie a questa soluzione si elimina la necessità di firewall est-ovest e di tecnologie obsolete e insicure come il Network Access Control, il quale contraddice i principi dell’approccio Zero Trust.

“Insieme, Zscaler e Airgap Networks hanno l’opportunità di reinventare ogni aspetto del modo in cui le soluzioni tradizionali hanno affrontato la sicurezza negli ambienti di campus e data center, in particolare a livello est-ovest. Zscaler sta puntando su SASE con un’innovazione significativa intorno a Zero Trust Networking, rivoluzionando SD-WAN, NAC e la sicurezza basata su firewall per le reti est-ovest e OT” ha concluso Kumar.

Condividi l’articolo



Articoli correlati

Altro in questa categoria


https://www.securityinfo.it/2024/05/03/zscaler-elimina-la-necessita-della-segmentazione-basata-su-firewall/?utm_source=rss&utm_medium=rss&utm_campaign=zscaler-elimina-la-necessita-della-segmentazione-basata-su-firewall




CISOs aren’t scapegoats: Fostering a security-first culture

Ten years ago, it was the norm for security breaches to be the sole responsibility of the chief information security officer (CISO). For this reason, the CISO role traditionally had a higher turnover rate, with many experiencing extreme burnout. But now, as data breaches make regular headlines and every organization becomes a lucrative target for cybercriminals, IT security has become a business priority, causing the full C-suite to take note. 

Cybersecurity is now a business enabler, directly impacting the bottom line with emerging regulations and making it more of a priority than ever for organizations. In fact, according to the SEC’s updated guidelines that went into effect at the end of last year, public companies now have four business days to report a cybersecurity breach that may impact an organization’s bottom line to the SEC. This means that CFOs and CISOs will have to work together to ensure breaches are disclosed on 10-K and 20-F forms. 

Security is about more than just a single moment in time. To keep up with today’s expanding attack surface and sophisticated threats, organizations must get proactive and create a culture of security that promotes a company-wide, always-on approach to assessing risks and defending against them. We’ve seen this evolution with financial and budget decisions, which each department now oversees, rather than leaving full responsibility to the CFO. Given the number of pathways that lead to a security incident, the culpability can no longer fall solely to the CISO — every department needs to have accountability. Here’s where organizations can start to foster such a proactive, company-wide approach to security that takes this into account:

Understand the threats of today

In today’s connected world, a company’s total number of attack surfaces and overall digital footprint continues to expand, giving adversaries many pathways to gain entry to networks and environments. As emerging technologies like artificial intelligence (AI) continue to infiltrate enterprises, the threat landscape gets even more complex — making it more challenging for cyber defenders to protect the business. In fact, a recent study found that 75% of security professionals witnessed an increase in attacks over the past 12 months, with an astonishing 85% attributing this rise to bad actors using generative AI.

In addition to the prevalent threat of generative AI, software supply chain security breaches are becoming increasingly alarming for many organizations, as attacks like MOVEit and Log4j resulted in millions of affected individuals and devices. These attacks are particularly concerning as it’s more difficult to understand what risks are brought in through third-party vendors. 

While external network risks rise, internal networks are even worse. A NetSPI report found that internal networks have nearly three times more exploitable vulnerabilities than external networks and that web applications have a higher prevalence of high and critical vulnerabilities compared to mobile and thick applications. Internal network vulnerabilities are further put at risk with generative AI technology. 

The expansion of the attack surface has also turned cybersecurity into an ‘all-hands-on-deck’ issue. Where the CISO used to be the sole head of cybersecurity, organizations now need a more holistic approach supported by the entire executive level. For example, CEOs and boards need to ensure the organization prioritizes security, CFOs need to ensure they are properly budgeting for security, and CIOs need to communicate with CISOs so the security team knows what applications are used that could pose a security risk. 

Foster a security-first culture from the top down

Security needs a holistic, proactive approach that involves the entire company doing its part to keep up with today’s security demands — from the CEO and CISO, down to the temporary intern and third-party vendors.  

With internal network vulnerabilities and web applications at the forefront of cyber professionals’ minds, developers and IT personnel must also consider keeping cybersecurity in mind during the development cycle. ‘Shifting left’ is not a new concept, but shifting security left is still a work in progress. However, as more and more internal network vulnerabilities crop up, it’s clear that application developers and security professionals need to communicate and collaborate to ensure end-to-end security. 

But security needs to go further than just creating and securing applications. Today, all employees and customers must have some cybersecurity knowledge. 

It only takes one employee to bring the whole system crashing down — reinforcing my earlier statement that an entire employee base needs to make security their responsibility, not just leave it up to their manager or CISO — making employee training critical. Employees should be trained from their first day to practice basic security hygiene and recognize phishing attempts; from there, training sessions should be regular and build off of the initial curriculum covered on day one. Training might not always prevent negligence, but it acts as an additional line of defense in the security arsenal, especially when done consistently. 

Additionally, organizations should provide cybersecurity training for their customers, or at the very least, ensure customers are informed about data breach procedures. While a customer falling for a phishing attempt won’t cause a company-wide data breach, organizations can still suffer reputational damage if a customer believes they are responsible for data theft. Providing customers with an organization’s data breach policies can prevent customers from falling for branded phishing or vishing attacks.

Going beyond the people, cybersecurity itself also needs a holistic and continuous approach. Even experts make mistakes — it’s impossible to ensure complete security, so continually testing security systems can provide a strengthened security posture.

Cybersecurity is rapidly changing. As a result, teams are seeing an evolution from the CISO as the head security practitioner to a leader who works across departments to oversee an organization’s security program. Cybercriminals are becoming more sophisticated, threats are evolving, and the attack surface is expanding. CISOs alone cannot prevent or even anticipate all potential threats, which is why it’s critical that the whole organization, as well as external partners, work towards creating a secure future.

https://www.securitymagazine.com/articles/100626-cisos-arent-scapegoats-fostering-a-security-first-culture




Report: The cost and complexity of data compliance impedes innovation

New research from Protegrity shows that organizations are utilizing data to promote innovation and create new strategies. However, less than 2% of respondents can access sensitive data within a week’s time. This suggests that a majority of organizations cannot completely utilize stored data. Yet, data remains a key component in many business operations, with 92% of respondents stating that data drives innovation with a positive impact on the bottom line, and all respondents stating that data can improve the experiences of both customers and employees. 

According to the report, 37% of businesses must wait one to two months in order to access sensitive data. 32% report waiting as much as three to six months.

The research also finds that organizations are following data regulations and compliance expectations. 

  • 45% use data controls to secure data. 
  • 84% are either somewhat or entirely prepared to meet PCI compliance expectations. 
  • 90% say that corporate executives understand the importance of data compliance and other regulations surrounding data. 

Data privacy and compliance is seen as a priority for many businesses, as 96% plan to invest part of the organization’s budget into data security. 49% plan to invest as much as 15% of the organization’s budget. 

https://www.securitymagazine.com/articles/100627-report-the-cost-and-complexity-of-data-compliance-impedes-innovation




Leveraging holistic GRC for compliance and audit preparation

Governance, risk, and compliance (GRC) is often treated as a separate entity from security. But compliance experts know that the two are hopelessly intertwined — after all, what’s the point of a security control if it isn’t directly tied to mitigating a risk? Regulatory frameworks and compliance standards provide externally vetted guidelines and internal governance standards provide a targeted measuring stick against which organizations can evaluate their security capabilities, allowing them to gauge their ability to defend against today’s most pressing threats. What’s more, they can help leaders on the business side of the enterprise better understand why certain solutions and policies are necessary.

But not all risk programs are created equal — and not every organization recognizes the critical role that GRC plays alongside departments like security, legal and business operations. Too many businesses remain unnecessarily siloed, with individual departments pursuing the goals they believe are best for them — without understanding how they fit into the organization as a whole. The approach might result in some short-term wins, but it’s rarely a recipe for long-term success. Organizations need to be able to take a step back and look at the bigger picture. A holistic approach to GRC can allow organizational leaders to better understand how their decisions will impact the organization’s overall risk profile — and, ultimately, determine whether those decisions are in the best interest of the business.

What holistic GRC means, and why it matters

Today’s organizations collect a significant amount of data, and they use that data to inform a wide array of business decisions. A lot of the data involves GRC — either directly or indirectly — but it also tends to be spread across multiple solutions, applications and platforms. Data access restrictions also limit who can work with that data, which can lead to communication silos where different departments have competing priorities that negatively impact the wider organization. Failing to aggregate that data so it can be viewed in the proper context means organizational leaders won’t see the big picture.

This creates inefficiencies within the organization. For example, the compliance team may believe they need a specific security solution in order to meet the requirements outlined in a certain framework. But they may not realize that the security team has already implemented a different solution that achieves the same result. Or they may not recognize that implementing their chosen solution will negatively impact the sales team by implementing cumbersome new data access restrictions. Holistic GRC isn’t just about looking at risks themselves — it’s about understanding what those risks mean for the business. That includes identifying the potential downstream effects that might result from addressing those risks. Purchasing a redundant security solution is likely a small problem. But grinding the sales team to a halt might be a bigger one.

What makes holistic GRC notable is that it can adapt to the organization, rather than the other way around. That’s important, because different organizations have radically different regulatory and compliance needs, and approaching a SOC 2 audit requires a different approach than, say, adhering to GDPR or CPRA data privacy regulations. But by approaching GRC in a holistic manner, organizations can better understand where their risk management program currently stands and how it will be affected by changes in the future.

Holistic GRC best practices

Gathering data is nothing new for organizations, but those that want to shift toward a more holistic approach to GRC can start by finding ways to integrate that data. It’s important to have a platform to view security controls, compliance needs, incident reports, policy information and other relevant information in a dynamic way. What organizations need more than anything is a way to view the relationships between different data sets in a responsive, real-time way, allowing them to better understand how changes within one area can affect the others. That solution might come in the form of a GRC platform, or the organization might choose a different, homegrown method of data integration — no matter how they accomplish it, the ability to easily visualize wide ranging data from across the organization is essential.

At its most basic level, understanding risk comes down to gauging how likely an event is to happen and what the potential impact of that event would be. Eliminating 100% of risk is impossible, but holistic GRC can help organizations better understand where their bases are well covered and where their most worrying gaps exist. This contextual data around the risks or gaps ensures that when communicating those risks upwards you have all the data to answer the hard questions you’ll inevitably receive. In addition, being able to map those risks to controls to policy statements to testing makes sure the risk mitigation plans are operating effectively. This is extremely helpful in circumstances like a SOC 2 audit. The SOC 2 framework is more concerned with outcomes than specific solutions, which means organizations enjoy a fair amount of latitude in terms of how they can approach compliance. As long as the organization can demonstrate how the spirit of the control maps to the letter of the control to the implementation, the auditor should be satisfied — which means the ability to clearly visualize and illustrate risk across the organization is extremely valuable. An effective holistic GRC program can help organizations demonstrate to auditors how they are approaching different aspects of compliance quickly and easily, without the need for cumbersome manual processes. 

On the flip side, holistic GRC can help organizations better understand what they need to do in order to comply with those frameworks and regulations that do have specific requirements. After achieving alignment with SOC 2 guidelines, a business may decide to enter a new international market that requires ISO 27001 certification. Because the two frameworks have significant overlap, an organization that leverages holistic GRC should have an easy time identifying which elements they are already in compliance with and which they still need to prioritize. The integration of security data, compliance data, cloud policies and other relevant information means organizations can more easily visualize where they stand in relation to different regulations, audit requirements and other benchmarks.

In addition to helping with compliance frameworks and audits, holistic GRC can serve as an important business enablement tool. At its core, GRC is about gauging risk — and understanding risk is at the core of making important business decisions. Plugging variables into a holistic GRC solution can help businesses better understand the downstream effects of certain actions or decisions. For example, it isn’t just important to know what steps the business needs to take to become ISO 27001 compliant — it’s important to know how much those steps will cost, and how that number stacks up against what the organization stands to gain and if/how that impacts the organization’s risk profile.

If achieving ISO 27001 compliance will be simple and the new market is a rich one, the decision should be relatively easy. On the other hand, if reaching (and maintaining) ISO 27001 compliance will take a substantial effort and the potential gains are marginal, the business might ultimately decide it isn’t worth the risk. Holistic GRC gives organizations the ability to consider compliance and security-based decisions in a business context, allowing them to make more informed decisions. That means limiting and avoiding risk, yes — but it also means understanding when some level of risk is worth the potential gain. Risk is a factor in every decision a business makes — but holistic GRC makes that risk easier to measure, understand and evaluate.

Using holistic GRC to speak the language of business

Perhaps the best thing about holistic GRC is the way it enables organizations to enjoy a real-time snapshot into their security and compliance posture, allowing them to more effectively measure how changes to their risk profile will affect the organization. Compliance audits and security evaluations are a fact of life for today’s businesses, and the ability to quickly and easily demonstrate adherence to different regulations and frameworks means the organization doesn’t have to invest significant resources and manpower into manual processes. What’s more, it allows the organization to evaluate its risk profile using the language of business, making it clear how different decisions impact not just one area of the business, but the organization as a whole.   

https://www.securitymagazine.com/articles/100625-leveraging-holistic-grc-for-compliance-and-audit-preparation




AI-enabled data collection and the regulatory landscape

As the world embarks on the Fourth Industrial Revolution, one major contributor is Artificial Intelligence (AI), specifically its ability to collect, process and analyze massive quantities of data at incredible speeds. Already, innovative enterprises are utilizing various AI tools and methodologies, such as web crawling, social media monitoring and digital analytics to anticipate emerging trends, deliver highly personalized customer services and boost productivity through finely tuned business operations. While it is undeniable that AI has elevated the value of data, the unprecedented level and sophistication of these collection activities have given rise to a new and growing regulatory landscape.  

The risks of using personally identifiable information

AI can analyze and utilize customers’ personally identifiable information (PII) in ways that could infringe on an individual’s privacy. A 2023 poll from Pew Research on Americans and data privacy reveals that consumers are wary of the implications of AI-driven data collection activities. In particular, 81% of those surveyed who expressed an awareness or knowledge of AI said that companies using AI to collect data would use it in a way that would make them uncomfortable. Eighty percent feared companies would use the collected data contrary to its original intent. Nevertheless, 61% admitted that personal data collected through AI could make their lives easier.

AI and PII concerns can fall into two categories: input and output concerns. The former category involves using large datasets and tools that contain PII. If the proper guardrails are not in place, there is a risk that AI-generated outputs could unintentionally expose sensitive information. And, as data becomes more valuable to enterprises, so too does it become more attractive to cyber criminals who benefit from stealing it. The latter concern deals with AI using PII to arrive at automated conclusions, such as denying someone access to a loan based on biased algorithms. Although this concern doesn’t specifically touch on data privacy per se, it does involve the misuse of PII.   

Prominent data privacy regulations in the European Union and United States  

Unsurprisingly, these data concerns surrounding AI and enhanced collection practices have increased regulatory scrutiny and compliance pressures across the globe. In the European Union, legislation is already in place in the form of the General Data Protection Regulation (GDPR) as well as the AI Act, which seeks to establish a common regulatory and legal framework for AI. Primarily, the GDPR focuses on protecting personal data, defining it as “any information relating to an identified or identifiable natural person.” Although the GDPR does not address AI explicitly, it contains many provisions relevant to AI.

In the United States, the National Institute of Standards and Technology issued the Risk Management Framework (AI RMF) last year, offering guidance to companies and other entities on using, designing and deploying AI systems. This framework is voluntary, and there are no penalties for noncompliance. However, there are privacy laws at the state level that carry penalties, including the Colorado Privacy Act, the Texas Data Privacy and Security Act and the California Privacy Rights Act (CCPA). The CCPA, in particular, was later amended to speak directly to AI with rules about “automated decision-making.”

Regulations for protected health information

The main data organizations will gather through AI is PII, but they could unconsciously collect more sensitive data categories like protected health information (PHI) and biometric data. PII encompasses the health-related information linked to an individual or disclosed during healthcare services. For example, health data is collected via period tracking apps, connected devices and genetics websites. AI-enabled collection tools could also gather data by analyzing customers’ social media posts, noting if they feel sick, depressed or happy. Likewise, AI tools like chatbots and facial recognition platforms collect and store PHI and biometric data.

Much like with PII, there are consequences for accidentally leaking PHI or ineffectively protecting it from bad actors. As such, the collection of PHI carries with it even stricter regulations — notably, the United States Health Insurance Portability and Accountability Act (HIPAA). Regulations, such as the Illinois Biometric Information Privacy Act (BIPA), aim to curb the indiscriminate collection of biometric data, including biological attributes like fingerprints or facial recognition patterns.

How companies can comply with these laws

Companies engaging in AI-powered data collection methods face a complex and ever-evolving regulatory landscape, where even unintentional noncompliance could result in serious fines and reputational damage. To effectively navigate these sensitivities, enterprises must build a comprehensive data governance framework, establish transparent and concise privacy policies and invest in thorough data mapping and inventory tools.

Businesses should ask for explicit consent to use sensitive data and set up rigorous data processing and storage protocols to not only comply with regulators but also help ease customer concerns. Organizations must also anonymize data where possible and restrict unwarranted data collection. Moreover, companies need to keep a pulse on the changing regulatory environment, noting those data protection laws relevant to their regions of operation.  

Using AI to improve data privacy 

Amid efforts to protect customers’ privacy, enterprises can find an unlikely ally in AI. Beyond harvesting and extracting insights for mountains of data, businesses can use AI algorithms to monitor data handling practices. Likewise, they can equip their customers with AI-enabled tools that allow them to easily request access to or erasure of their data. AI-powered detection systems can also spot potential cybersecurity threats for proactive prevention. However businesses opt to use AI, maintaining compliance will be indispensable with the diverse and evolving regulatory landscape.

https://www.securitymagazine.com/articles/100628-ai-enabled-data-collection-and-the-regulatory-landscape




Gli LLM non riescono (ancora) a sfruttare le vulnerabilità in autonomia


Sempre più spesso si discute delle crescenti capacità degli modelli di IA e dei rischi che derivano da usi scorretti, anche in ambito cybersecurity. Molti ricercatori di sicurezza si chiedono fino a che punto gli LLM possono supportare le attività del cybercrimine, per esempio creando campagne malware su larga scala o sfruttando vulnerabilità in autonomia, senza l’intervento umano.

Secondo Chris Rohlf, ricercatore di sicurezza, i modelli di IA non sono ancora in grado di scoprire e generare nuovi exploit per le vulnerabilità note dei sistemi in completa autonomia, ma solo di sfruttare la conoscenza esistente per automatizzare il processo. 

Per illustrare la sua posizione, Rohlf analizza un paper in cui i ricercatori dimostrano come un LLM sia stato capace di sfruttare quasi tutte le CVE usate nel test; nel dettaglio, il gruppo ha sviluppato costruito un agente in grado di interfacciarsi con GPT-4 e altri modelli e ha testato le sue capacità di exploit con una lista di 15 vulnerabilità note. Le CVE sono tutte presenti in software open source, citate molti altri paper accademici e, cosa più importante, sono state scoperte dopo il cut-off di conoscenza di GPT-4 per valutare se effettivamente l’agente stava “ragionando” sulla vulnerabilità e non semplicemente usando le informazioni che aveva già a disposizione.

Il test dei ricercatori ha dimostrato che GPT-4 è riuscito a sfruttare l’87% delle vulnerabilità semplicemente avendo accesso alla sua descrizione; in assenza di essa, il tasso di successo è sceso al 7%; al contrario, gli altri modelli non sono riusciti a sfruttare alcun bug pur avendo accesso alle stesse informazioni di GPT-4.

Visti i risultati, gli autori del paper concludono che il modello ha dimostrato la capacità emergente di sfruttare autonomamente vulnerabilità 1-day reali. Secondo Rohlf, però, si tratta di una conclusione affrettata che non riflette le reali abilità degli LLM.

Pixabay

Gli LLM non sono così bravi a sfruttare le vulnerabilità

Il ricercatore evidenzia innanzitutto che gli autori del paper non hanno rilasciato né i prompt utilizzati, né il codice dell’agente, né gli output del modello per evitare rischi di sicurezza, e ciò rende il loro processo poco trasparente e difficile da approfondire. 

Rohlf spiega inoltre che l’agente illustrato nel paper presenta un modulo capace di effettuare ricerche sul web, di fatto in grado di trovare tutte le informazioni tecniche per sfruttare le CVE note. Gli exploit delle vulnerabilità sono semplici da reperire e da eseguire, composti solo da qualche linea di codice, e sono accessibili direttamente dal National Vulnerability Database. “In molti casi il link NVD è il primo risultato di ricerca su Google” afferma Rohlf.

L’agente basato su LLM non ha quindi dimostrato capacità innovative di ragionamento, ma ha semplicemente usato le abilità che aveva per effettuare ricerche sul web ed eseguire gli exploit esistenti delle vulnerabilità; nulla che un qualsiasi agente con moduli di web scraping non possa fare.

“La mia conclusione è che questa ricerca sta dimostrando la capacità di GPT-4 di essere utilizzato come scanner e crawler intelligente che si basa ancora su alcuni approcci di forza bruta anche una volta ottenuti i giusti passaggi di sfruttamento, e non una capacità emergente di cybersecurity” spiega Rohlf.

Si tratta di un caso d’uso legittimo e che può effettivamente velocizzare l’esecuzione di exploit, ma in questo caso non si può parlare di “hacking autonomo” o di generazione automatica di exploit. GPT-4 non sta scoprendo nuove vulnerabilità né sta generando nuove modalità di sfruttamento, ma sta semplicemente raccogliendo le informazioni già disponibili sul web per fornirle all’utente.

“La pubblicazione di ricerche di questo tipo, senza i dati a sostegno delle affermazioni, può rafforzare la falsa idea che i modelli di IA siano pericolosi per la sicurezza informatica e debbano essere controllati. Questo non è vero per lo stato dell’arte attuale dei modelli” afferma Rohlf.

Pur non trattandosi di una nuova capacità degli LLM, vale comunque la pena sottolineare che questi chatbot hanno la reale capacità di velocizzare i processi manuali e facilitare l’esecuzione di exploit anche a chi non ha conoscenze tecniche. D’altra parte, l’accesso a un grande volume di informazioni in poco tempo è un vantaggio significativo anche per chi si occupa della difesa dei sistemi.

Gli LLM hanno un enorme potenziale di migliorare la cybersecurity più di quanto non possa supportare il cybercrimine, ma, sostiene Rohlf, c’è ancora un grosso gap tra gli esperti di intelligenza artificiale e cybersecurity; di conseguenza, l’impressione è di combattere contro una minaccia molto più grossa di ciò che è in realtà.

Condividi l’articolo



Articoli correlati

Altro in questa categoria


https://www.securityinfo.it/2024/05/03/gli-llm-non-riescono-ancora-a-sfruttare-le-vulnerabilita-in-autonomia/?utm_source=rss&utm_medium=rss&utm_campaign=gli-llm-non-riescono-ancora-a-sfruttare-le-vulnerabilita-in-autonomia




Gli U.S.A. pubblicano nuove linee guida per gestire i rischi dell’IA


Il Dipartimento della sicurezza interna degli Stati Uniti (Department of Homeland Security – DHS) ha pubblicato nuove linee guida per proteggere le infrastrutture critiche del Paese e le armi di distruzioni di massa dai rischi dell’IA.

“L’intelligenza artificiale può offrire soluzioni trasformative per le infrastrutture critiche statunitensi, ma comporta anche il rischio di rendere questi sistemi vulnerabili a guasti critici con nuove modalità, attacchi fisici e informatici. Il nostro Dipartimento sta prendendo provvedimenti per identificare e mitigare queste minacce” ha affermato Alejandro N. Mayorkas, segretario del DHS.

Pixabay

Le nuove linee guida contro i rischi dell’IA

Insieme al CISA, il DHS ha rilasciato delle linee guida cross-industriali volte a ridurre l’impatto dei rischi legati all’IA e proteggere i sistemi delle infrastrutture critiche.

Le due organizzazioni hanno lavorato seguendo tre diverse categorie di rischio: la prima contempla l’uso dell’IA per pianificare, migliorare o scalare attacchi fisici o cyber alle infrastrutture; la seconda considera gli attacchi contro i sistemi di IA che supportano le infrastrutture critiche; infine, nell’ultima categoria rientrano tutti i rischi legati a errori presenti nell’implementazione o l’esecuzione dei tool di IA.

La strategia si organizza in quattro punti fondamentali che responsabili e utenti dell’infrastruttura devono seguire per mitigare i rischi. In primo luogo, le linee guide impongono di creare una cultura improntata sulla gestione del rischio dell’IA; ciò si traduce nel dare priorità alle implicazioni di sicurezza, alla trasparenza dei sistemi e rendere la protezione una priorità di business.

Le imprese devono inoltre definire il proprio profilo di rischio individuando i contesti in cui possono emergere e sviluppare sistemi che consentono di analizzare e tracciare i rischi dell’IA e gli impatti.

Infine, è indispensabile implementare e mantenere sistemi di controllo dei rischi per massimizzare i benefici dei sistemi di IA e al contempo ridurre la probabilità di impatti negativi.

intelligenza artificiale

Pixabay

Proteggere le armi di distruzione di massa

Oltre alle linee guida per proteggere i sistemi critici, il DHS, insieme al Countering Weapons of Mass Destruction (CWMD) Office ha pubblicato un report che analizza il rischio che l’IA venga usata per supportare lo sviluppo e la produzione di minacce chimiche, biologiche, radioattive e nucleari. Nel documento il Dipartimento ha inserito anche una serie di step da seguire per mitigare la diffusione di queste potenziali minacce contro la nazione.

“L’uso responsabile dell’IA è molto promettente per far progredire la scienza, risolvere sfide urgenti e future e migliorare la nostra sicurezza nazionale, ma  richiede anche di essere preparati a mitigare rapidamente l’uso improprio dell’IA nello sviluppo di minacce chimiche e biologiche” ha affermato Mary Ellen Callahan, sottosegretario del CWMD.

Le pubblicazioni arrivano dopo 180 giorni dall’ordine esecutivo di Biden sulla gestione sicura e responsabile dei sistemi di intelligenza artificiale e mirano a rafforzare la sicurezza di organizzazioni e cittadini. Come per l’AI Act europeo, anche in questo caso si parla di indicazioni piuttosto generiche che ogni organizzazione deve poi implementare contestualmente, con tutte le difficoltà che comporta. Il successo delle regolamentazioni statunitensi si basa sull’idea di una stretta collaborazione tra aziende e istituzioni; senza di essa, gli sforzi potrebbero risultare vani.

Condividi l’articolo



Articoli correlati

Altro in questa categoria


https://www.securityinfo.it/2024/05/02/gli-u-s-a-pubblicano-nuove-linee-guida-per-gestire-i-rischi-dell-ia/?utm_source=rss&utm_medium=rss&utm_campaign=gli-u-s-a-pubblicano-nuove-linee-guida-per-gestire-i-rischi-dell-ia