AIaaS (7)

The AI that Learned how to Cheat and Hide Data from it's Creators


The AI that learned how to cheat and hide data from its creators

  • AI was trained to transform aerial images into street maps and then back again
  • They found that details omitted in final image reappeared when it was reverted
  • It used steganography to 'hide' data in the image and recover the original photo

New research from Stanford and Google has shown that it's possible artificial intelligence software may be getting too clever. The neural network, called CycleGAN, was trained to transform aerial images into street maps, then back into aerial images. Researchers were surprised when they discovered that details omitted in the final product reappeared when they told the AI to revert back to the original image.


Stanford and Google researchers were surprised when they discovered that details omitted in the final product reappeared when they told the AI to revert back to the original image. For example, skylights on a roof that were absent from the final product suddenly reappeared when they returned to the original image, according to TechCrunch.  

'CycleGAN learns to "hide" information about a source image into the images it generates in a nearly imperceptible, high-frequency signal,' the study states. 'This trick ensures that the generator can recover the original sample and thus satisfy the cyclic consistency requirement, while the generated image remains realistic.'

What ended up happening is that the AI figured out how to replicate details in a map by picking up on the subtle changes in color that the human eye can't detect, but that the computer can pick up on, TechCrunch noted. In effect, it didn't learn how to create a copy of the map from scratch, it just replicated the features of the original into the noise patterns of the other. 

For example, skylights on a roof that were absent from the aerial reconstruction suddenly reappeared when they returned to the original image, or the aerial photograph labeled (a)

Researchers found the AI figured out how to replicate details in a map by picking up on the subtle changes in color that the human eye can't detect, but that the computer can pick up on. The researchers say the AI ended up being a 'master of steganography,' or the practice of encoding data in images. CycleGAN was able to pick up information from the original source map and then encode it in the reconstructed image. By doing that, it enables the AI to be able to recover the original image with precise accuracy. However, it means that the AI was using steganography to avoid actually learning how to perform the requested task in order to speed up the process, TechCrunch noted.


AI systems rely on artificial neural networks (ANNs), which try to simulate the way the brain works in order to learn. ANNs can be trained to recognise patterns in information - including speech, text data, or visual images - and are the basis for a large number of the developments in AI over recent years.

Conventional AI uses input to 'teach' an algorithm about a particular subject by feeding it massive amounts of information.

AI systems rely on artificial neural networks (ANNs), which try to simulate the way the brain works in order to learn. ANNs can be trained to recognise patterns in information - including speech, text data, or visual images. Practical applications include Google's language translation services, Facebook's facial recognition software and Snapchat's image altering live filters. The process of inputting this data can be extremely time consuming, and is limited to one type of knowledge. 

A new breed of ANNs called Adversarial Neural Networks pits the wits of two AI bots against each other, which allows them to learn from each other. This approach is designed to speed up the process of learning, as well as refining the output created by AI systems. 

Share this story

Why Your Business Needs to Embrace AI if You Don’t Want to Be Left Behind

The machines are taking over, it’s true. But you don’t have to run and hide just yet. At the moment, artificial intelligence is still run and managed by humans. And billions of us all over the world use it every day of our working lives.

You may be a business owner who has heard of AI and its benefits for business use. But like 45% of US respondents surveyed for Computer Weekly, you may not feel ready to embrace machine intelligence within your current business infrastructure.

The purpose of this article is to explore the relationship that your business needs to have with AI. Among the issues we consider are: your business goals, the price of AI, how it can help your business get ahead of your competition, and security.

You have a unique set of business goals

The first thing that you should consider is where AI fits into your company’s unique business goals, as for each goal there will be a range of different AI enhancement options.

This is to ensure you are bringing the relevant qualities AI offers to the areas of your business that will benefit from the technology. For instance, you may want to consider updating your operations to:

 1. Increase customer service capabilities

AI chatbots can help homepage visitors complete transactions without having to speak to a real-life person. You can also use AI to collect customer service feedback and improve your team’s productivity. AI-led customer service tools can even make it possible for customers to track their deliveries or raise query tickets, without human intervention. Self-service technologies like order tracking take some of the responsibility for mundane tasks away from your staff.

 2. Reduce inefficiencies in your supply chain

AI in your supply chain means that your solutions and frameworks will be constantly improving themselves and developing over time. The best way to use AI is to enable autonomous action – like having AI-assisted machines monitor POS data and make predictions about future purchase habits and consumption trends. This kind of real-time data could have a major positive kickback from a scalability and roll-out perspective.

 3. Get better ROI from your marketing campaigns

You may have  have poured lots of time and money into marketing campaigns, yet you still can’t seem to garner a regular stream of audience interaction. If this is so you may want to consider how AI can help you.

AI bots are able to draw from a deeper range of source data than traditional marketing research techniques. This allows you to be more refined in your method of targeting your customers and increase the engagement and response levels you see from your campaigns  – David Steinberg, the co-founder and CEO of Zeta Global, has claimed that marketing campaigns that incorporate AI have an ROI that is up to 1600% higher than those which do not.

You can find AI at every price point

You don’t have to shell out on Amazon warehouse style robotic systems to reap the full benefits of AI within your business.

Why Your Business Needs to Embrace AI if You Don’t Want to Be Left Behind

In many instances, just ensuring that you incorporate the basic AI technology that is relevant to your business and industry will be enough to help you stay in the loop.

For ecommerce brands, the shopping cart service you choose will be vital to scaling your brand’s growth. Make an online store with access to a continually developing list of AI functionality, such as marketing coaching tips.

You can also access higher-priced AI tools such as beacon technology for brick and mortar retail outlets. These tools help brands create immersive advertising campaigns that send connected mobile users push notifications with incentivizing offers.

The options and scale with which a brand can utilize AI can be staggering and awe-inspiring. While this may seem daunting, it is important for you to remember that, whatever your budget is, you must make sure that you get to grips with AI.

AI can help you develop unique ideas to beat the competition

AI-assisted tools are helping retail brands stay ahead of the competition by ensuring that their stores look lovely at all times. Shelfie cameras attached to shop shelving provide live feedback data to staff’s mobiles. Here, they can be notified when an item is missing or misplaced within the store.

The addition of RFID tags can also be used within retail to create a unique and immersive shopping experiences. Burberry, for instance, offers shoppers a ‘magic mirror’ that will recommend clothing to customers in fitting rooms.

Further, IBM created E.L.F for Mall of America in the run-up to Christmas 2016. By logging in through Facebook Messenger, mobile users in the mall could access ELF’s services, offering suggestions for customers searching for gifts for their loved ones.

AI will be at the forefront of security

As internet payment and technologies expand and increase, so will the need for tighter security against data breaches. Analyzing and protecting your customer’s data is an integral part of building trust as a business. AI can help your business track buying behavior and alert you to any areas of unusual activity within your business operations.

Using deep learning and NLP Processing such as Aphelion's Singularity, AI is becoming all the more sophisticated in recognizing and alerting to possible cyber threats. While this is, of course, very useful to your business, you will also need to make sure that investigations into threats are lead by your data team.

This is to ensure that there is accountability for your AI, so that you are certain it is accurately recording threats. In addition to this, having your data team overseeing the process means they can draw lessons from the threats your AI notes, meaning that you can continue to improve the range and level of security offered by your business

However, without AI intervention, in the years to come it may become harder to stay on top of the hundreds of ‘strikes’ that may befall your company’s systems daily.

So, there you have the reasons why your company needs to embrace AI to stay ahead of the game. You can start off small and gradually scale up your AI offerings to build a highly efficient network of systems.

The future is bright for AI, so make sure you invest if you don’t want to be left behind. To discuss and explore the possibilities how AI could benefit your business Contact Aphelion.


Share this story

The Role of AI in Cybersecurity


The growing and evolving cyber security risk facing global businesses can be stemmed by the integration of AI into security systems 

 The Role of AI in Cybersecurity

Hyper-connected workplaces and the growth of cloud and mobile technologies have sparked a chain reaction when it comes to security risks. The vast volume of connected devices feeding into networks provide a dream scenario for cyber criminals — new and plentiful access points to target. Further, security on these access points is often deficient.

For businesses, the desire to leverage IoT is tempered by the latest mega breach or DDoS attack creating splashy headlines and causing concern.

However, the convenience and automation IoT affords means it isn’t an ephemeral trend. Businesses need to look to new technologies, like AI, to effectively protect their customers as they broaden their perimeter.

The question becomes, how can enterprises work with, and not against, artificial intelligence?

>See also: How AI has created an arms race in the battle against cybercrime

The emergence of AI in cyber security

Machine learning and artificial intelligence (AI) are being applied more broadly across industries and applications than ever before as computing power, data collection and storage capabilities increase. This vast trove of data is valuable fodder for AI, which can process and analyse everything captured to understand new trends and details.

For cyber security, this means new exploits and weaknesses can quickly be identified and analysed to help mitigate further attacks. It has the ability to take some of the pressure off human security “colleagues.” They are alerted when an action is needed, but also can spend their time working on more creative, fruitful endeavours.

A useful analogy is to think about the best security professional in your organisation. If you use this star employee to train your machine learning and artificial intelligence programs, the AI will be as smart as your star employee.

Now, if you take the time to train your machine learning and artificial intelligence programs with your 10 best employees, the outcome will be a solution that is as smart as your 10 best employees put together. And AI never takes a sick day.

It becomes a game of scale and leveraging these new tools can give enterprises the upper hand.

AI under attack

AI is by no means a cyber security panacea. When pitted directly against a human opponent, with clear circumvention goals, AI can be defeated. This doesn’t mean we shouldn’t use AI, it means we should understand its limitations.

AI cannot be left to its own devices. It needs human interaction and “training” in AI-speak to continue to learn and improve, correcting for false positives and cyber criminal innovations.

This hybrid approach already has proven itself to be a valuable asset in IT departments because it works efficiently alongside threat researchers.

Instead of highly talented personnel spending time on repetitive and mundane tasks, the machine takes away this burden and allows them to get on with the more challenging task of finding new and complex threats.

Predictive analytics will build on this by giving security teams the predictive insight needed to stop threats before they become an issue as opposed to reacting to a problem. This approach is not only more cost effective in terms of resources, but also is favourable for the business due to the huge reputational and financial damage a breach can cause in the long term.

Benefits of machine learning

Alongside AI, machine learning is becoming a vital tool in a threat hunter’s tool box. There is no doubt machine learning has become more sophisticated in the past couple of years and will continue to do so as its learnings are compounded and computing power increases.

Organisations face millions of threats each day, so it would be impossible for threat researchers to analyse and categorise them all. As each threat is analysed by the machine, it learns and improves. This not only helps protect organisations now, but compiles this valuable data for use in predictive analytics.

However, just staying ahead of the hackers and the threats they pose is not enough to protect organisations as the new vulnerabilities and new devices that come online will make this more and more difficult.

The continued and enhanced standardisation on data formats and communication standards is crucial to this effort. Once data flows and formats are clearly defined, not just technically but also semantically, machine learning systems will be far better placed to effectively police the operations of such systems.

The industry needs to work towards finding the sweet-spot between unsupervised and supervised machine learning so that we can fully benefit from our knowledge of current threat types and vectors and combine that with the ability to detect new attacks and uncover new vulnerabilities.

Much like AI, machine learning in threat hunting must be guided by humans. Human researchers are able to look beyond the anomalies that the machine may pick up and put context around the security situation to decide if a suspected attack is truly taking place.

The future

For the security industry to get the most out of AI, they need to recognise what machines do best and what people do best. Advances in AI can provide new tools for threat hunters, helping them protect new devices and networks even before a threat is classified by a human researcher.

Machine learning techniques such as unsupervised learning and continuous retraining can keep us ahead of the cyber criminals. However, hackers aren’t resting on their laurels. Let’s give our threat researchers the time to creatively think about the next attack vector while enhancing their abilities with machines.

Share this story

How AI has Created an Arms Race in the Battle Against Cybercrime


The growing capabilities of artificial intelligence is triggering a battle across the cyber security fence – and organisations must act now to be on the right side of it 

How AI has Created an Arms Race in the Battle Against Cybercrime

Artificial intelligence (AI) has been increasing in sophistication for some years, finding its place in our everyday lives with ever-growing pace and force. As businesses and governments begin to use AI, the potential for its application in cyber security is becoming more apparent.

What’s more, hackers and businesses are going head-to-head – with hackers now able to develop more sophisticated threats, and businesses looking to use AI for threat detection, prevention and remedy.

When it comes to cyber security, businesses need to act now to tighten up cyber defences. With large-scale security breaches only increasing in number over recent years, organisations both big and small should consider investing in AI systems designed to bolster their defences.

>See also: The rise of the machine: AI, the future of security

Over the next year alone, we’ll see a rise in AI systems that can perform several tasks, including re-writing encryption keys continuously, preventing them from being unlocked by hackers outside of an organisation’s walls.

These more practical uses for AI are allowing organisations to anticipate issues before they arise through threat analysis, threat detection and threat modelling. For example, if a human manually checked systems for signs of outside breaches on a monthly basis, it could take a number of weeks to fully analyse. Using AI not only adds an extra layer of protection, but also allows organisations to react to the breach much quicker.

Hackers will up their AI game

Vulnerabilities found both in software and online have previously been numerous, offering hackers plenty of opportunities without great need for AI. This will quickly change as AI improves and businesses minimise the gaps within their organisation’s cyber defences.

It may not be long before the use of AI becomes the norm among hackers, providing them with more opportunities and avenues to access sensitive data. This technology could be used to scan the internet and software for vulnerabilities, as well as design attack strategies, and then launch them with minimal human error.

One current use of AI by cybercriminals is in phishing emails. By using data from the target to send phishing emails that replicate human mannerisms and content, these AI-powered attacks resonate with the target better than ever before. These tactics will make it harder for businesses and individuals to recognise when they’re being hacked.

Tackling insider threats

Of course, many threats to an organisation originate far closer to home. Insider threats have always been a cause for concern, but as the potential of AI systems grow in complexity, we are starting to see businesses tackle this with force.

AI can now help to detect a break from normal employee behaviour. This technology could be used to discover employees that are accessing company information, and evidence of them transferring this information outside of organisation walls.

Taking this to a more invasive level, AI technology could be used to detect instances of corporate policy being breached by employees. Tasks as harmless as using USB storage can now be analysed for signs of malicious intent and corporate corruption.

Of course, exact sentiment and explanation will be difficult to detect from AI technology alone. As a result, privacy laws will be key if organisations are to avoid breaches in employee law themselves.

Skills gap

Keeping the ball in the court of the cyber security teams will be an increasingly hard battle to fight in the coming years, and one which will need the full support and expertise of cyber security professionals and security-savvy organisations.

With the Centre for Cyber Safety and Education revealing that the world will face a shortfall of 1.8 million cyber security professionals by 2022, we are reaching a critical point where change is needed rapidly.

This is something that has been recognised by the government in recent months, with announcements made in the Budget demonstrating a commitment to address the skills shortage.

The introduction of T-Levels will aid in the creation of the next generation of technology professionals, helping to fill the widening gap in provision and part of this must focus on cyber security.

As the nature and complexity of AI grows, businesses need to start thinking about how to incorporate this new technology into their cyber security strategies. Of course, not everyone is a target for such advanced AI attacks and simple cyber hygiene remains an effective counter to many threats.

However, there is plenty of evidence that AI is becoming more available and affordable and so will become more prevalent. But if organisations are to truly take advantage, a combined effort is needed.

Not only must organisations invest in preventative AI, but the government must continue to back the development of the next generation of technology professionals. After all, there’s no use in having the technology without professionals that know how to use it.

Share this story

Bring the noise: How AI can improve cyber security


‘Researchers are now modelling how a malevolent AI system could develop, and have concluded that current cyber security practices are woefully inadequate’ 

Bring the noise: How AI can improve cyber security

Beleaguered enterprises are struggling to keep pace with cyber threats, and small and medium-sized businesses are hit hardest of all due to limited resources.

A recent survey by the Federation of Small Business (FSB) found 66% of those questioned had been a victim of cybercrime over the past two years, and only 4% had an incident response plan in place in anticipation of an attack.

For many, cyber security takes them into unfamiliar territory and depletes the time spent on core business activities.

This has seen an over-reliance upon point solutions, poor attention to patching and updates, and a failure to apply strategic business-specific security controls.

To make matters worse, the potential attack surface is only set to widen as the Internet of Things sees sensors and IP-enabled tech insinuate themselves into every niche of society, even the small business.

A badly configured humble kettle could open up a conduit onto a business network, for instance. Yet the current situation finds many SMEs ill-prepared for any change in the threat spectrum, being unable to monitor, detect and respond to an attack – begging the question, how will they cope with yet more holes in the network?

What is needed is some form of automation coupled with artificial intelligence; a system that has visibility of the network and can monitor activity and alert the business to enable security resources to be focused as and where needed, thereby conserving spend, but which is specific to the business.

High-level data processing has been available for some time in the form of security incident and event management (SIEM) systems that, when combined with a security operations centre (SOC), can correlate data and issue alerts.

But these systems can be costly and complex to deploy and manage, with reports estimating it takes up to six personnel to run a SOC 24/7.

Even then, the information derived from these tools needs to be correctly interpreted and actioned upon. And few SMEs have data scientists on the pay roll.

For this reason, AI is beginning to receive more attention. It takes complex event processing and performs pattern analyses, using machine learning, to improve success rates.

In the context of a SOC, AI can be used to extract hidden correlations and detect complex attack vectors, as well as by assisting analysts looking for traditional attack patterns by offering multiple filtering options.

It can then assess the potential for these events to scale-up and evolve into attacks. Threat feeds are assessed in the context of the business, so that criteria such as geography, sector and compliance requirements are used as parameters externally, while internal elements, such as business strategy and the risk profile, are included to create an overarching view –allowing the threat to be assessed against the risk appetite of the business before determining a response.

As opposed to a traditional SOC, an AI SOC demonstrates machine learning and uses deep threat intelligence. It can drill down further for data and use advanced penetrative techniques to mine information from dynamic data sources such as those associated with social media and even off-grid in the dark web.

This can give the business advance warning of an impending attack in real-time as data can be collated, sifted and interpreted using predictive data analytics to forecast likely event outcomes.

The FSB survey found that the most common form of attack against the SME were phishing attacks experienced by 49% of respondents, with 37% experiencing the more targeted spear phishing attack.

These can readily be spotted and filtered using automated software. Trickier and more difficult to anticipate are denial of service attacks, aimed at crippling websites, and ransomware attacks, which use DDoS attacks or malware to demand a release fee.

Both are on the increase in the SME sector, with the FSB survey reporting five percent of respondents had experienced a DoS attack and 4% ransomware.

By the time a DoS has been executed, the business is already caught off guard and is potentially in a capacity war, forced to scale up resource to fend off the attack.

Yet, with sufficient warning, the SME can use a DoS solution to throttle the attack. The key is getting that information in advance for it to become actionable intelligence and that can only be achieved by applying AI in the form of complex algorithms that can spot rogue activity.

For instance, DoS attacks are highly organised in nature and are often planned on forums hosted on the dark web. Tap into those conversations by using the parameters referred to above and you can create a window into underground activity that can trigger an alert when the noise merits it.

Real-time SOC services are now emerging that can deliver this type of capability to the SME and it doesn’t need to cost. Outsourcing can provide the SME with access to the technology, the AI, and the personnel needed to man the operation, thereby giving the sector access to high-level security services using economies of scale for the first time.

When selecting a supplier, it’s the intelligence that you need to look for, so in addition to the usual requirements such as SIEM, event logging and data analytics, it’s beneficial to look at the managed services on offer.

Ask how data is captured and correlated and analysed and by whom? Can it dovetail with your day-to-day business operations to provide business intelligence?

Finally, bear in mind that the threat spectrum is constantly evolving. Cyber security sees security solutions and attackers pitted against one another in a never-ending arms race.

If we now have AI security solutions, businesses should expect to see malicious AI systems in the future.

Researchers are now modelling how a malevolent AI system could develop, and have concluded that current cyber security practices are woefully inadequate.

Share this story

The Rise of the Machine: AI and the Future of Security


The need for a cyber security overhaul is necessary as IT professionals know signature matching is no longer an effective means to identifying current attacks. 

The Rise of the Machine: AI and the Future of Security

AI has impacted our day-to-day lives for years, whether that’s automated voice calls or virtual personal assistants – like Siri – or even self-driving cars.

The next step is to implement AI technology into personal and cyber security systems.

Currently, one or two guards will monitor a bank of security screens, and it is a successful method of security, but it is not full proof.

Eliminating human error is a key driver behind bringing Artificial Intelligence to security through intelligent video analytics.

Humans can easily get distracted, generally have short attention spans, and often find it difficult to focus on multiple things at once – a bank of security screens.

In an article written by Dr. Mahesh Saptharishi, Senior Vice President of Analytics and Data Science at Avigilon, he explains: 'While a security officer might miss a person sneaking into a poorly lit facility, a camera backed with intelligent video analytics is designed to catch a flash on the screen and recognize it as a potential threat.'

'It will spot a person loitering at the perimeter of a schoolyard and alert on-the-ground security officials to investigate and take action if necessary, all without missing a beat and keeping close watch on the many cameras and locations.'

Just as AI can be applied to personal security systems, so to can it with cyber security systems.

The need for a cyber security overhaul is necessary as IT professionals know signature matching is no longer an effective means to identifying current attacks.

Hackers can easily conceal their attacks from these signature matching security systems.

A rejuvenation of the current system is needed.

>See also: Bring the noise: How AI can improve cyber security

Yesterday, DB Networks announced its DBN-6300 and Layer 7 Database Sensor software, were being deployed to successfully implement AI in the cyber security environment – to automatically protect databases’ infrastructure.

"AI-based cyber security is truly a sea change in the security industry," said DB Networks' Chairman and CEO Brett Helm. "AI enables us to quickly and accurately…identify cyber attacks in progress. In future generations of product, DB Networks will use the output from AI to drive autonomous cyber security technologies that not only block attacks but also automatically heal the vulnerabilities."

Caution, as always, must be taken – a Skynet scenario (for those of you who have seen Terminator), while unlikely, is not beyond the realm of possibility given this is the direction human technology is heading.

But the integration of AI into personal and cyber security systems is a natural progression as technology develops. It is more efficient and not hindered by human error.

Artificial Intelligence will of make locations – physically and virtually – safer by making technology more efficient and adaptable.

Share this story

How MSPs Can Streamline AI Development for their Customers


Many companies are looking at starting AI development projects to assess the potential of AI and Machine Learning technologies in their operations, but getting set up for AI and ML development can be a daunting task as it means integrating software and gaining access to GPU processors that cost thousands of dollars. By offering AI-as-a-service to their customers, MSPs can take the risk, hassle, and much of the cost out of starting on the path to AI development.

AI products are quickly becoming commonplace, and AI applications and solutions are now more viable than ever with the availability of modern machine learning and deep learning tools such as TensorFlow and Keras, along with GPUs that are built specifically to perform parallel operations on large amounts of data. Enterprise applications for AI run the gamut from health sciences to finance, security, data center operations and cyber surveillance, and companies are eager to try these applications to improve agility, reduce costs and improve efficiency.

General-purpose CPUs cannot deliver the user responsiveness and inference latency required by complex deep learning and AI workloads. Instead, these new workloads require the dedicated horsepower of GPUs that were designed for them. The problem is that building a GPU-based AI development capability is complex and expensive, and companies may not want to spend tens of thousands of dollars and hundreds of person-hours just getting set up to begin development. Also, one needs to have a shared cluster where GPU resources can be allocated to end users on demand and taken back once the project completes.

Aphelion's AI-as-a-service offering makes it possible for MSPs to take the risk and hassle out of getting started with AI development. AI-as-a-service takes care of managing GPU resources distributed across a set of hosts in a multi-tenant manner plus all of the OS and CUDA library dependencies, so users can focus on AI development. MSPs can also give their users dedicated access to multiple GPU resources without making them invest in their own GPU platforms. Furthermore, one can automate deployment of an applications and software development platform for AI using pre-installed AI and machine learning software-based images and application templates. This provides single-click deployment of software development and machine learning environments for end users.

At Aphelion, we believe that that the key to differentiation and profits for MSPs is in offering customized, white-glove services that the big cloud companies can’t offer, and we offer a simple, cost-effective way to deliver those services. Aphelion makes it easy for MSPs to offer popular services with a series of service templates that run on our Intelligent Cloud Platform, including AI-as-a-service, DevOps-as-a-service, GPU-as-a-service, and VPN-as-a-service. Aphelion will continue to expand its group of click-and-go service engines to ensure that MSPs have the best possible suite of services that drive revenues, profits and happy customers.

Share this story
Subscribe to this RSS feed