Home Blog Page 71

India selects Sarvam AI to develop first AI foundational model

  • Planned model, envisioned to be capable of reasoning, voice integration, and fluency in multiple Indian languages, promises to unlock a plethora of opportunities.

The Indian government’s selection of Sarvam AI, a Bengaluru-based startup, to spearhead the development of the nation’s first indigenous AI foundational model marks a significant milestone in India’s technological trajectory.

The initiative, announced alongside the Electronic Components and Manufacturing Scheme (ECMS) guidelines by IT Minister Ashwini Vaishnaw, underscores the government’s commitment to fostering a robust and self-reliant AI ecosystem within the country.

Sarvam AI’s selection from a pool of 67 proposals highlights the company’s promising approach and the potential for impactful contributions to India’s AI landscape.

Holistic approach

Founded in July 2023 by Vivek Raghavan and Pratyush Kumar, both alumni of AI4Bharat, Sarvam AI is poised to develop a comprehensive suite of generative AI solutions.

Their focus encompasses the entire spectrum of AI development, from cutting-edge research in custom model training to the creation of a sophisticated, enterprise-grade platform for AI model authoring and deployment.

This holistic approach is crucial for ensuring that the developed foundational model is not only technologically advanced but also readily accessible and adaptable for diverse applications across various sectors.

The planned model, envisioned to be capable of reasoning, voice integration, and fluency in multiple Indian languages, promises to unlock a plethora of opportunities.

Bridging the digital divide

By prioritising these capabilities, Sarvam AI aims to address the unique needs of the Indian populace and bridge the digital divide by enabling AI-powered solutions that are accessible and relevant to a wider range of users.

The emphasis on secure and large-scale deployment further solidifies the model’s potential for widespread adoption and integration into critical national infrastructure.

Sarvam AI’s co-founders, Vivek Raghavan and Pratyush Kumar, have articulated a clear vision for the project, emphasising collaboration with Indian enterprises to co-develop domain-specific AI models.

This collaborative approach, coupled with the commitment to build, deploy, and optimise the model within India using local infrastructure and talent, underscores a commitment to national ownership and control over this critical technology.

This aligns perfectly with the IndiaAI Mission’s broader objective of promoting strategic autonomy and enhancing domestic innovation in the AI domain.

Furthermore, Sarvam AI’s collaboration with AI4Bharat at IIT Madras, a leading institution in Indian language AI research, demonstrates a commitment to leveraging existing expertise and fostering synergistic partnerships within the Indian AI community.

The development of three model variants – Sarvam-Large, Sarvam-Small, and Sarvam-Edge – suggests a strategic approach to catering to a diverse range of computational requirements, from advanced reasoning to real-time interactive applications and on-device tasks.

Key steps for data exfiltration prevention in mobile devices

  • It is important for enterprises to put in place appropriate mobile device management strategies for the protection of sensitive, proprietary information against unauthorised access and data breaches.
  • The approach that is proactive or may mean prioritising upholding prevention of data exfiltration in itself is a justification for protecting corporate assets and gaining trust and assurance in mobile technology as a core component of modern business processes.

In the present age, with data being the most precious asset, appropriate MDM practices cannot just be an option but a dire necessity for data security.

Mobile devices have played increasingly crucial roles in personal as well as professional situations in today’s digital environment. With dependence on mobile technology rising, organisations have to find ways of preventing the theft of sensitive data-in other words, the malicious transport of information from within an organisation to an outside entity.

It is by effective mobile device management (MDM) that sensitive information can be protected from unauthorised access and compliance with data protection regulations ensured.

Mobile device management (MDM) refers to a set of solutions and tools used to manage and monitor corporate employees’ mobile devices when they use to connect to corporate resources. This kind of solution provides IT administrators with very powerful capabilities, including inventory for each device, enforcement of configurations, application management, and security of data.

Prevent data exfiltration within the organisation using these essential strategies that organisations can implement in their MDM practices.

  1.  Strong Authentication Methods:  Unless a strong authentication process is followed, mobile devices cannot be sufficiently safeguarded. Organisations should implement two-factor authentication systems that require users to present more than one proof before giving access to devices and applications. Biometric methods, like fingerprints or facial recognition, should be complemented by other traditional forms, like passwords and PINs, to further ensure denial of entry.
  2. Encryption of Data:  Unauthorised access to mobile data can be prevented by ensuring that it is encrypted. By doing so, even if someone gains access to the data, it becomes unreadable in the absence of an appropriate decryption key. This becomes another safety barrier thrown across the pathway of data breaches. Both at rest and during transit, data should be encrypted using strong encryption standards and all sensitive data should be well protected.
  3. Application Whitelisting or Blacklisting:  Application whitelisting policy enables control by an organisation of what applications can be allowed to run on devices, which thus minimises risks associated with malicious software or rogue apps that aid in data exfiltration. Otherwise, it also further blocks access to potentially malicious software by maintaining a blacklist of known harmful applications.
  4. Regular Security Audits and Monitoring:  Conduct regular security audits to identify weaknesses and misconfigurations in the organisation’s MDM. It also serves the organisation well if continuous monitoring of mobile devices is done for suspicious activities such as unauthorised access attempts or unusual data transmissions, as it provides a quick response to what could turn out to be a data breach. Automated alerts will facilitate prompt investigation and remediation.
  5. User Education and Awareness Training: Employees are usually the first line of defence against data exfiltration. Therefore, it is critical to provide them with extensive training and education on best practices for mobile-device usage, including recognition and management of phishing attacks and the dangers of public Wi-Fi. Regularly conducted awareness campaigns increase the personal commitment to security measures and the practices that encourage responsible behaviour.
  6. Remote Wipe Functionality:  If a device is lost or stolen, then remote wipe capability is important for the organisation. Organisations should adopt an MDM solution that can support remote deletion in order to ensure that confidential information does not remain in the wrong hands if it is put on the wrong devices. This becomes increasingly relevant for organisations that work with sensitive information or PII.
  7. Laws and Regulations Compliance: Last but not least, the mobile device management procedure should comply with all the laws and regulations standards, such as the General Data Protection Regulation (GDPR) or the Health Insurance Portability and Accountability Act (HIPAA). Compliance is not only beneficial for the protection of organisational data but also acts as a confidence booster among its clients and stakeholders.

Data exfiltration prevention examples

Data breaches could occur through cyberattacks, insider threats, and even accidental leaks. Attackers could exploit the vulnerability in a system or network to siphon off confidential information, while employees may take advantage of their clearances to extract sensitive data for malicious purposes, or even unknowingly share it through poor security practices.

The costs associated with data exfiltration could be severe, including financial loss and operational downtime, not to mention reputational damage and legal liability.

  1.  Data Loss Prevention (DLP) Solutions: DLP technology features are predominantly meant to identify, supervise and protect important data that may take residence at end-user points, network endpoints, and cloud services. Such a deployment may include DLP programs that trigger keyword or pattern matching of outgoing emails to detect Social Security numbers or credit card information. DLP systems have greatly reduced the chances of accidental and malicious data breaches because messages could be automatically blocked or encrypted.
  2. Network Security Measures: Robust network security measures should be in place for data exfiltration prevention-fortified firewalls, intrusion detection and prevention systems (IDPS), and secure web gateways which have traffic monitoring and access control. For instance, an organisation may well configure its firewall to block communication with known blacklisted IP addresses connected to data theft or limit outbound traffic to only necessary protocols and destinations.
  3. Endpoint Protection: Laptops and mobile devices usually serve as endpoints and, to some degree, they are targets for data exfiltration. Organisations may also protect themselves from endpoint attacks through endpoint protection platforms (EPP) combined with advanced threat detection and response capabilities. Enhanced security may be achieved by multi-factor authentication, which requires several proof-of-identity components before accessing data. Limiting the use of USB ports to an organisation could also prevent unauthorised transfer of files from devices directly.
  4. User Behaviour Analytics (UBA): UBA solutions leverage machine learning into the analysis of user behaviour besides detecting anomalies of possible data exfiltration attempts. For example, if an employee downloads hundreds of megabytes of information out of normal hours or accesses sensitive information that does not apply to their position, then automated alerts can notify security to further investigate. Such proactive methods would be able to help organisations quickly respond to potential illegal activities.
  5. Data Encryption:  Encryption of required data, both at rest and while in transit, is fundamental towards preventing data exfiltration. Following this, an intruder cannot decrypt an encrypted file without its appropriate decryption key. An organisation may, for instance, use Transport Layer Security (TLS) as an encrypted path for data transmission over the internet. It prevents even interception of the transmitted information.

How to build a strong defence against credential theft prevention?

  • As organisations increasingly have reliance on mobile devices for day-to-day processes, one of the important components is implementing a robust Mobile Device Management policy.
  • The dynamics keep changing every day, and therefore, proactive action against credential theft will be significant in securing sensitive information and trust from customers and stakeholders.

Credential theft is, by far, one of the slyest and most invasive forms of cyber threats. It involves unauthorised access to asymptotic vital data like usernames and even passwords with the end goal of unauthorised access to systems and data.

So, how can organisations put in place mechanisms to build defences against these threatening trends?

The state of credential theft

Credential theft can occur through various methods, such as phishing attacks, malware distributions, social engineering, and password cracking.

According to cybersecurity reports, phishing is one of the most common techniques used by cybercriminals – when an unsuspecting user is tricked into divulging their login credentials through deceptive emails or messages purporting to be from legitimate sources; the other could include keystroke logging malware, accessing security vulnerabilities.

The most shocking statistic of all is that well over 80 per cent of data breaches involve compromised credentials. In recent times, several high-profile incidents have made the case for the theft of credentials. Some breaches have exposed millions of user accounts while financially draining those corporations and government agencies that incurred regulatory penalties because of security breaches.

Organisations should therefore understand that credential theft has adverse domino effects, disrupting immediate assets as well as trust amongst customers, partners, and stakeholders.

Strategies for prevention

As this concern can take a serious turn when it comes to credential theft, organisations will have to stretch their hands in all directions for appropriate possible prevention techniques. It is all about technology, training users, and commitment to changing practices. Organisations are increasingly incorporating stringent measures, such as Mobile Device Management (MDM), to address this escalating concern.

  1.  Strong Password Policy:  Among all the preventive measures regarding credential theft, strong password policies are one of the strongest. Organisations should use complex passwords that span a mix of letters, numbers, and symbols with at least 12 characters as a standard in configurable password policies. They should also include features like frequent password changes and a prohibition on reusing passwords. Educating employees on the relationship between the importance of uniqueness might further strengthen this barrier against unauthorised access.
  2. Provide Multi-Factor Authentication (MFA): MFA greatly magnifies security since multiple factors for validation must be passed before access to accounts may be granted. The combination of something the user knows: a password; something the user possesses: a mobile device for the verification code; and something through-and-through: biometrics like fingerprints gives MFA a heavy embargo against credential theft. When it comes to cybercriminals that are heavily concentrating on the default password-based systems, MFA implementation renders it a game-changer for preventing unauthorised access to systems.
  3. Security Awareness Training: Employees make up the weakest link in the security defences of an organisation. Thus, they need education on the various tactics that are used in cybercrimes. Some organisations, for example, educate their employees on regular security awareness courses covering all of these schemes-such as phishing and social engineering schemes. Role-playing scenarios and simulated phishing attacks can make training much more effective by keeping employees on their toes and preventing real-life scenarios from occurring. Here, for instance, is where a culture of awareness and responsibility where employees feel responsible for guarding their credentials will contribute to an overall protected organisation.
  4. Regular Access Monitoring and Auditing: The best way an organisation can prevent the stealing of credentials is to put strict access control measures within a system. Organisations should regularly audit user access and make changes to the permissions using the principle of least privilege-giving rights only to what they need. Continuously monitoring log-in attempts and accounting for any anomalies or unusual access patterns will help organisations identify those already compromised accounts and respond quickly to minimise damage.
  5. Investment in Advanced Technologies:  Apart from compulsory user training and policy implementation, organisations must seek further advanced security technologies. Security Information and Event Management (SIEM) systems provide centralized security monitoring and real-time analytics and alerts for suspicious activities. Threat intelligence applications will help in making organisations ahead of known and emerging threats by enabling them with real-time information about malicious actors and their tactics. Endpoint detection and response (EDR) offers organisations an endpoint level tool with which to combat malware and other attacks.

Role of robust Mobile Device Management

It is necessary to enforce comprehensive security practices owing to an increase in mobile device usage for corporate purposes. A robust MDM (Mobile Device Management) service will act as a critical solution in fortifying an organisation against these threats.

A strong MDM service is critical to organisations wishing to improve their security posture. Organisations can secure sensitive data and systems against the constantly changing landscape of cyber threats by utilising multi-factor authentication, advanced MDM features, including remote wipe and geofencing, and centralised device oversight.

MDM solutions deliver analytics and alerts compliant with all devices, which help in swift remedial action against possible security threats. Keeping all devices up-to-date with security patches and updates can further strengthen organisations from exploitation.

Centralized monitoring and management of devices are also essential elements of keeping the corporate systems intact.

Endpoint security as a frontline defence

Endpoint security is the process through which endpoints are being secured-these are devices such as computers, smartphones, and tablet devices that connect to corporate networks.

They are mainly the entry points that cybercriminals want to compromise or exploit. Typically, without a well-established security strategy-the organisation is open to a range of threats including phishing attacks, malware, and unauthorized access opportunities.

Each of these would, to a certain extent, compromise sensitive authentication credentials and lead to a lot of financial and reputational loss.

Robust endpoint security measures can help an organisation proactively block unauthorised access and minimise risk from credential theft. Such measures include deploying antivirus software, endpoint detection and response (EDR) solutions, firewalls, and encryption technologies.

Furthermore, they should ensure regular updating of system users and training as prorated knowledgeable employees will recognize and eliminate potential threats.

Benefits of Remote Browser Isolation in safeguarding internet

  • Prospect of integrating Remote Browser Isolation improves opportunities for better security management and regulatory compliance.
  • The security improvements in using RBI will yield lots of benefit in terms of defense enhancing against threats from other parts of the web, reducing IT management overhead, and providing an excellent end-user experience.
  • By making the most of these technologies, organisations can help enable mobility without sacrificing agility and productivity against the increasingly unpredictable demands of today’s organisations.

The inevitable cyber space offers a lot of business-related threats such as computer viruses, ransomware, and phishing. Sensitive information has become really unsafe; hence the integrity of the whole organisation is at stake. There is growing evidence of this phenomenon as it raises awareness that Remote Browser Isolation (RBI) is a paradigm shift in safe browsing.

Handling advanced devices can do anything at the end of the day-from accessing important company data to doing transactions. In this case, isolation becomes even more important with regards to the ability of remote browser isolation because coupled with mobile device management (MDM), it demonstrates plenty of benefits to companies wanting to improve their security posture.

What is RBI?

Remote Browser Isolation is a security technology that removes the web browsing function from the end-user’s device and executes it in a cloud-based environment. Formerly, the browser isolates web content from a local device and allows the rendering to a secure server. It’s a different kind of approach but gives a higher layer of protection against web-related threats.

In addition, one of the most outstanding advantages of Remote Browser Isolation is better protection.

Benefits of RBI

RBI minimises the risk of infection associated with downloading applications from phishing websites and other cyber threats, which could put very sensitive data in jeopardy. MDMs, as we know, are already there to supplement in enforcement and monitoring device compliance with the security policies of various organisations. The role this will play in RBI will be in the layered defense against web threats.

Secondly, RBI improves the aspect of data privacy as web browsing is a significant leakage point. With isolation in the whole process of browsing, RBI closes any downloaded files or data from being completely opened to unauthorised applications on the device.

The integrated use with MDM will keep the compliance parameters like the General Data Protection Regulation (GDPR) and Health Insurance Portability and Accountability Act (HIPAA) across all the mobile devices. Compliance breaches or data misuse will be minimised.

On top of this, Remote Browser Isolation is actually going to enhance user experience while ensuring security. Without a doubt, users can surf the internet without worrying because RBI keeps unsafe web browsing activity separate from their workplace tasks.

This creates an environment where usage of company devices becomes a culture and improves productivity since employees can hardly experience interruptions which have come about due to security alerts or even slowdowns by malicious content on their devices.

Also, centralised management of all browsing activities is another added advantage by allowing IT admins to manage devices as well as compliance from a single dashboard of MDM Solutions. But onboarded with it, you would enable visibility around the very browsing behavior at real time, thus seeking to understand the possible threats or even be proactive in responding to newly identified threats.

This is needed for any enterprise serious in defining the controls it has over its security posture.

Last, but not least, the cost-effectiveness of adopting remote browser isolation along with MDM should also be included in the equation. This can be used to drive down expensive remediation costs and minimise the impacts and frequency of security incidents, thereby reducing downtime. Also, RBI reduces the need to have a strong arsenal of endpoint security tools as it provides a solid first line of defense against threats originating from the World Wide Web for most situations.

Commercial RBI vendors

  1. Menlo Security: Menlo Security is a pioneer organisation in the RBI arena. It offers solutions that block malware from reaching endpoints by running all content in the cloud. This architecture minimises data exfiltration chances and maintains user productivity thanks to an intuitive interface.
  2. Symantec (Broadcom): Symantec has created their Web Isolation solution that is currently targeting users into RBI technology to protect them against harmful online content. This would altogether work with the full and integrated portfolio of services that Symantec offers in-house, for free endpoint protection, threat intelligence, etc.
  3. Zscaler: The RBI services of Zscaler work on its structure built for cloud-security, thus creating zero-trust architecture. It directly helps secure access to the Internet and cloud applications while isolating all possible threats away from the user’s device to make it truly a seamless browsing experience.
  4. Cato Networks: Among the range of its offerings, Cato Networks includes RBI as part of its all-encompassing secure access service edge (SASE) platform. This integration allows organisations to use remote browser isolation along with their other security methods, ensuring that the traffic is filtered and the malicious content is destroyed before reaching the endpoint.
  5. Forcepoint: Forcepoint’s RBI solution is one of its purview safety offerings. With a strong focus on user behavior and data safety, it successfully isolates and analyses web content while providing user context.

Open-source RBI solutions

  1. Web Isolation Framework (WIF): An open-source project that enables the deployment of an RBI solution at the user end, this framework provides the backbone for building secure browsing environments and can be extended with other features upon demand.
  2. Sandboxing Technologies: Technologies like Firejail or Google Sandboxie can be leveraged to build isolated environments for web browsing. These tools help in narrowing down the attack surface by confining potentially damaging web apps to a well-defined environment.
  3. Qubes OS: Although not an RBI-only tool, Qubes OS is an open-source operating system, employing virtualisation to create compartments in which applications, including web browsers, are separated. Thus, risks are minimised by constraining any cyber threats within a specific qube while online.
  4. Browser Sandbox. Some developers have produced browser extensions for the popular web browsers that enhance their isolation features. Although perhaps not being totally equivalent to the complete isolation provided by dedicated RBI solutions, these extensions do help reduce risk when browsing.

Hackers, rather than hacking, steal devices to gain sensitive data

  • It is an easier path for cybercriminals compared to navigating complex digital security protocols.
  • Stolen devices often lack robust data protection measures, making them vulnerable to exploitation.

While cybersecurity threats like sophisticated firewall breaches garner significant attention and investment, a recent study by Kensington highlights the persistent and often underestimated risk posed by physical device theft.

The study reveals a concerning reality: stealing a laptop, desktop, tablet, or external storage device can be a surprisingly effective method for accessing sensitive data, representing a tangible and often easier path for cybercriminals compared to navigating complex digital security protocols.

This enduring vulnerability demands renewed attention from organisations, particularly given the potential for significant financial, legal, and reputational repercussions.

Ancillary concerns

The Kensington study, surveying 1,000 IT decision-makers across various industries, paints a stark picture. A staggering 76 per cent of respondents had dealt with device theft in the past two years, demonstrating the widespread nature of the problem.

The immediate consequences extend beyond the mere loss of hardware, with data breaches identified as the paramount concern by 46 per cent of participants. This fear is well-founded, as stolen devices often lack robust data protection measures, making them vulnerable to exploitation.

The resulting unauthorised access to company data, cited by 43 per cent of respondents, further exacerbates the potential for damage.

Moreover, the survey highlights ancillary concerns such as visual hacking (23 per cent) and the vulnerability of sensitive data on insecure home networks (23 per cent), emphasising the interconnected nature of security risks in the modern, mobile workforce.

The impact of device theft extends far beyond abstract security concerns, translating into tangible financial and operational consequences.

Economic burden

The study reveals that 33 per cent of organisations experienced legal or regulatory consequences due to compromised data stemming from stolen devices. Furthermore, 32 per cent reported disruptions to employee productivity, while 30 per cent faced financial losses associated with device replacement and increased insurance costs.

These figures underscore the significant economic burden imposed by physical security breaches, often dwarfing the perceived cost of implementing preventative measures.

The escalating cost of data breaches, as highlighted by the IBM Cost of Data Breach Report 2024, further amplifies the gravity of the situation. With the global average cost of a data breach reaching $4.88 million, the potential liability associated with a stolen device containing sensitive information is considerable.

Ironically, despite recognising the importance of physical security solutions, the Kensington study found that 60 per cent of organisations were not utilising security locks for their devices. These disconnect between awareness and implementation reveals a critical gap in organisational security strategies.

While reports from organisations like Kroll indicate that data theft losses have surpassed physical losses since 2010, dismissing the latter as insignificant would be a grave oversight.

Physical theft remains a persistent and often simpler avenue for cybercriminals seeking access to sensitive data. It bypasses complex cybersecurity measures like firewalls and intrusion detection systems, relying instead on the inherent vulnerability of unsecured devices.

WhatsApp beefs up its chat privacy feature

  • By disabling the export of chats and preventing automatic media downloads, WhatsApp is effectively creating a walled garden around its conversations.
  • Future iterations, as indicated by WhatsApp’s announcement, will likely address remaining vulnerabilities and further empower users to manage their digital footprint.

WhatsApp, a leading messaging platform, has introduced “Advanced Chat Privacy,” a feature designed to bolster user discretion and prevent the unauthorised dissemination of content. WhatsApp’s commitment to privacy is not new.

The platform has long implemented end-to-end encryption, ensuring that messages remain confidential between sender and recipient. Moreover, the introduction of features like disappearing messages, which prevent forwarding, copying, saving, or screenshotting of media, demonstrates a continuous effort to enhance user control over their shared content.

Advanced Chat Privacy represents a further refinement of this approach, directly addressing concerns about the export of chats, automatic media downloads, and the potential misuse of messages for AI training.

Creating a walled garden

By disabling the export of chats and preventing automatic media downloads, WhatsApp is effectively creating a walled garden around its conversations. This is particularly relevant in group chat environments where individuals may not fully trust all participants.

The ability to restrict the use of messages for AI features speaks to the growing anxieties surrounding data harvesting and the potential for sensitive information to be utilised for unintended purposes.

The cumulative effect of these restrictions is a heightened sense of security and control for the user, potentially encouraging more candid and open communication within the WhatsApp ecosystem.

However, it is important to acknowledge the limitations of Advanced Chat Privacy. While it provides a significant barrier against unauthorised data extraction, it cannot entirely eliminate the risk of malicious activity. Users should remain vigilant about the information they share, even with the added protection this feature offers.

Furthermore, the fact that this is only the first iteration of the feature suggests that WhatsApp acknowledges the need for further development and refinement. Future iterations, as indicated by WhatsApp’s announcement, will likely address remaining vulnerabilities and further empower users to manage their digital footprint.

Related Posts:

Meta’s AI integration in WhatsApp ignites wave of users’ discontent