Archive for July 31, 2024

Palo Alto Networks Rolls Out Secure AI by Design Portfolio

Posted in Commentary with tags on July 31, 2024 by itnerd

What: Palo Alto Networks is making available to customers its Secure AI by Design product portfolio, aiming to secure organizations’ GenAI usage and development of enterprise AI applications by providing visibility, control, and protection specific to AI, addressing new risks and threats. As businesses increasingly integrate AI, the portfolio enables them to confidently build and use AI-powered apps, while also prioritizing the integrity of AI security frameworks from development to deployment.

Why: The need for securing AI applications has become increasingly important as businesses continue to integrate AI and LLMs into their operations. With employees adopting AI applications at a rapid pace and organizations across various industries gaining a competitive edge through AI-powered applications, the Secure AI by Design portfolio aims to securely enable AI deployments.

While the promises of AI are significant, it’s essential to acknowledge the associated risks with equal emphasis in order to realize its full potential. Bad actors are using AI to ramp up the scale of attacks, so it is important that organizations are proactive in their defense.

How: Organizations will be equipped to create a secure AI ecosystem that prioritizes the integrity of AI security frameworks from development to deployment. Businesses can fully harness the potential of AI without compromising security through the following use cases:

  • Securely enable GenAI applications: With the growing trend of employees using GenAI apps for business purposes, AI Access Security enables organizations to use AI tools with confidence. It gives security teams full visibility, application and data access controls, and continuous data risk monitoring.
  • Fortify AI supply chain: Businesses must be aware and rectify against possible risks. With Prisma Cloud AI Security Posture Management (AI-SPM), organizations can secure their AI ecosystem by identifying vulnerabilities and misconfigurations in models, applications and resources. It improves compliance and minimizes data exposure, thus improving the integrity of your AI security framework.
  • Protect enterprise AI applications: It is critical for organizations to see every component of their AI app ecosystem— including AI applications, models, inference and training datasets. AI Runtime Security is designed to help solve this, and protect against evolving zero-day and AI-specific threats, such as data leakage from AI models and applications, and safeguard models from misuse and attacks.

When: To start the roll out, AI Runtime Security is now available on Google Cloud and will be available later in August on Amazon Web Services (AWS) and Microsoft Azure. To follow, AI-SPM will be available on August 6 and AI Access Security will be available on August 19.

Additional Information: Learn more about our Secure AI by Design portfolio, read our latest blogs on AI Runtime Security and AI-SPM. Explore Precision AI by Palo Alto Networks, which powers our cybersecurity platforms and solutions.

Introducing Mission AI Foundation

Posted in Commentary with tags on July 31, 2024 by itnerd

Mission Cloud today announced the launch of Mission AI Foundation, a comprehensive service designed to help businesses optimize their AI solutions on AWS while adhering to best practices and managing costs effectively.

83% of companies claim that AI is a top priority in their business plans. However, this widespread interest in AI adoption is often tempered by significant challenges. Concerns such as financial management, security, and knowledge gaps frequently stand in the way of companies eager to harness AI’s power to accelerate innovation and drive process efficiency.

To address these barriers, Mission AI Foundation combines support, engineering, strategy, and guidance to help organizations manage their cloud infrastructure and build for the future of AI.

The service is built on five pillars of continuous engagement:

  • AI solutions optimization, architecture guidance, and prompt engineering
  • Continuous cost optimization and ongoing financial management
  • Guidance on foundational best practices and cloud governance
  • Round-the-clock Enterprise support, led by Mission Cloud and backed by AWS
  • Cloud strategy for adopting technologies that transform your business

Key features of Mission AI Foundation include:

  • Access to a team of certified Cloud Analysts, Technical Account Managers, Solutions Architects, and AI Engineers
  • Improving token usage to ensure AI operations are economical and high-performing
  • Engineer Assist – AI, offering pay-as-you-go engineering support for AI solutions
  • Continuous guidance on prompting best practices to improve accuracy, optimize template, reduce hallucinations, and enhance overall model performance
  • 24/7 support with AWS Enterprise-level SLAs
  • Large Language Model Operations (LLMOps) to build and maintain your dedicated operations pipeline

Mission AI Foundation leverages Mission Control, the company’s cloud services platform, along with Amazon QuickSight and Amazon Q, to provide detailed cost visualization and management tools. The service also includes carbon footprint tracking, offering insights into the environmental impact of AI workloads.

Mission AI Foundation is now available. For more information or to schedule a demo, click here or contact sales@missioncloud.com.

Study Shows How Inclusive Social Media Really Is

Posted in Commentary with tags on July 31, 2024 by itnerd

The WizCase Team has conducted some research with the aim to understand “How inclusive is social media really?” by creating fresh social media accounts, and showing differences between platforms before algorithms kick in and take over.

According to their research:

  • White people were represented more than any other group — ranging from 51-60% of all posts.
  • TikTok had the most gender parity, with a 50/50 representation, and YouTube had the least, with 68% males and 32% females.
  • Only 4 posts in our sample from X contained any political messaging, 100% of those were conservative.

Why it matters:

Increasingly, fears of and reactions to “cancel culture” and censorship on social media make headlines. Conservatives claim they are being silenced, while liberals claim that hard-right accounts are being inappropriately boosted. Both sides are saying that more representation is needed for their group.

You can access their detailed report here: https://www.wizcase.com/blog/how-inclusive-is-social-media

Microsoft Says Yesterday’s Outage Was Caused By A DDoS Attack

Posted in Commentary with tags on July 31, 2024 by itnerd

Yesterday, Microsoft had a major outage with a lot of their online services. And today, Microsoft is serving up a reason for that outage:

An unexpected usage spike resulted in Azure Front Door (AFD) and Azure Content Delivery Network (CDN) components performing below acceptable thresholds, leading to intermittent errors, timeout, and latency spikes. While the initial trigger event was a Distributed Denial-of-Service (DDoS) attack, which activated our DDoS protection mechanisms, initial investigations suggest that an error in the implementation of our defenses amplified the impact of the attack rather than mitigating it.

So Microsoft was hit by a DDoS attack, and their automatic defences amplified it rather than mitigating it. At least they were honest about it and came out with this statement quickly. You can be sure that Microsoft will correct the issues that they found here so that a DDoS attack doesn’t have the same effect as it did yesterday.

OMG! $75 Million Ransom reportedly Paid To The Dark Angels Ransomware Group

Posted in Commentary with tags on July 31, 2024 by itnerd

ZScaler researchers have revealed that an undisclosed victim paid a record $75 million ransom to the Dark Angels ransomware group earlier this year. This figure nearly doubles the previous highest ransom of $40 million, paid by CNA Financial in 2021.

The Dark Angels group, reportedly a rebranded version of the Babuk ransomware family, first emerged in May 2022. The group’s most notorious attack occurred in September 2023, targeting Johnson Controls, an automation and manufacturing company. SentinelOne reported that the attackers “used Dark Angels ransomware to lock the company’s VMWare ESXi servers” and demanded $51 million, though it remains unclear if the ransom was paid.

Steve Hahn, Executive VP, BullWall had this to say:

“Unlike other forms of cyber attacks, Ransomware has ties to Russia in virtually all instances. Whether it’s Russian based groups or Russia based tools or hacking services, nearly every attack has deep Russian ties. The proxy war we are engaged in via Ukraine has amped up Russia’s attacks on “soft targets” in the US that will do the most damage, along with getting the most ransom. Out of control inflation in the US over the last 4 years can be amplified by attacks on the supply chain or production. This, along with the fact that manufacturing can be slightly behind on their security journey, makes them a very easy and obvious target. 

Virtually all Ransomware attacks include a double extortion technique. The threat actors will encrypt the data and steal the data. You’ll pay to decrypt (to get your data back) and you’ll pay again to not have them leak embarrassing information or sell your secrets on the dark web. In addition, virtually all Ransomware attacks are preceded by a command and control element. Threat actors slowly get in to the Network, then use special tools like Cobalt Strike to get admin credentials, with admin credentials they can do whatever they desire. Including encrypting your backup or disabling your security products. 

All of these groups are connected to Russian Hackers. The groups and affiliations merge and morph. They use different encryption payloads but most are nearly undetectable by modern EDRs if they can launch them in the customers environment.

Total Ransomware payments have exploded to over a billion dollars a year, in just known payments, the number of successful attacks has more than doubled in the last 2 years and the average payment has gone up from $850,000 to over $2M in 2024. 

Organizations need defense in depth. Layers and layers of security. They also need to focus on not just preventing Ransomware, but also containment and recovery. They need to treat a successful Ransomware attack as a “when” not “if” scenario and plan accordingly. Table stakes is also MFA, Microsgementation and Zero Trust. But these are silver bullets. They still need to figure out how they react and respond quickly when that event occurs. 

Zero Trust is important. It will certainly lessen the chances of an attack. But this journey is typically very long for customers and still not a silver bullet. Zero day attacks, Shadow IT, personal devices, IoT devices, these are all attack vectors for Ransomware and once the encryption begins at the shares drives, whether those are cloud or local, it’s only a matter of time before all of the data is encrypted, even with zero trust network architecture in place. 

And of course, AI will increase the volume of phishing attacks and allow each and every email a custom look. AI tools can search all public information about their “target”, LinkedIn, Yelp, Facebook, Venmo payments. Anything public to find recipients to interact with and form a custom phishing email to the user. This means the number of highly targeted attacks will explode in the coming years. Along with that, AI can create Ransomware variants in milliseconds that will appear to EDR and antivirus as something they’ve never seen, making it incredibly difficult for them to stop those attacks. Along with that, dark web brokers are connecting people with footholds and access in companies to those with the payloads. The Facebook Marketplace for hackers.”

There’s a lot to unpack from Mr. Hahn, but it is very much worth reading. Organizations need to be prepared for all eventualities. As in keeping the bad guys out, and what to do if they get in. And most importantly, they should never, ever pay up as that simply encourages these threat actors. Frankly, your best defence is to do everything possible to not be a victim.