Recommender systems are AI-based algorithms that use different information to recommend products to customers. We can say that recommender systems are a subtype of machine learning because the algorithms “learn from their past,” i.e., use past data to predict the future.

Today, we’re exposed to vast amounts of information. The internet is overflowing with data on virtually any topic. Recommender systems are like filters that analyze the data and offer the users (you) only relevant information. Since what’s relevant to you may not interest someone else, these systems use unique criteria to provide the best results to everyone.

In this article, we’ll dig deep into recommender systems and discuss their types, applications, and challenges.

Types of Recommender Systems

Learning more about the types of recommender systems will help you understand their purpose.

Content-Based Filtering

With content-based filtering, it’s all about the features of a particular item. Algorithms pick up on specific characteristics to recommend a similar item to the user (you). Of course, the starting point is your previous actions and/or feedback.

Sounds too abstract, doesn’t it? Let’s explain it through a real-life example: movies. Suppose you’ve subscribed to a streaming platform and watched The Notebook (a romance/drama starring Ryan Gosling and Rachel McAdams). Algorithms will sniff around to investigate this movie’s properties:

  • Genre
  • Actors
  • Reviews
  • Title

Then, algorithms will suggest what to watch next and display movies with similar features. For example, you may find A Walk to Remember on your list (because it belongs to the same genre and is based on a book by the same author). But you may also see La La Land on the list (although it’s not the same genre and isn’t based on a book, it stars Ryan Gosling).

Some of the advantages of this type are:

  • It only needs data from a specific user, not a whole group.
  • It’s ideal for those who have interests that don’t fall into the mainstream category.

A potential drawback is:

  • It recommends only similar items, so users can’t really expand their interests.

Collaborative Filtering

In this case, users’ preferences and past behaviors “collaborate” with one another, and algorithms use these similarities to recommend items. We have two types of collaborative filtering: user-user and item-item.

User-User Collaborative Filtering

The main idea behind this type of recommender system is that people with similar interests and past purchases are likely to make similar selections in the future. Unlike the previous type, the focus here isn’t just on only one user but a whole group.

Collaborative filtering is popular in e-commerce, with a famous example being Amazon. It analyzes the customers’ profiles and reviews and offers recommended products using that data.

The main advantages of user-user collaborative filtering are:

  • It allows users to explore new interests and stay in the loop with trends.
  • It doesn’t need information about the specific characteristics of an item.

The biggest disadvantage is:

  • It can be overwhelmed by data volume and offer poor results.

Item-Item Collaborative Filtering

If you were ever wondering how Amazon knows you want a mint green protective case for the phone you just ordered, the answer is item-item collaborative filtering. Amazon invented this type of filtering back in 1998. With it, the e-commerce platform can make quick product suggestions and let users purchase them with ease. Here, the focus isn’t on similarities between users but between products.

Some of the advantages of item-item collaborative filtering are:

  • It doesn’t require information about the user.
  • It encourages users to purchase more products.

The main drawback is:

  • It can suffer from a decrease in performance when there’s a vast amount of data.

Hybrid Recommender Systems

As we’ve seen, both collaborative and content-based filtering have their advantages and drawbacks. Experts designed hybrid recommender systems that grab the best of both worlds. They overcome the problems behind collaborative and content-based filtering and offer better performance.

With hybrid recommender systems, algorithms take into account different factors:

  • Users’ preferences
  • Users’ past purchases
  • Users’ product ratings
  • Similarities between items
  • Current trends

A classic example of a hybrid recommender system is Netflix. Here, you’ll see the recommended content based on the TV shows and movies you’ve already watched. You can also discover content that users with similar interests enjoy and can see what’s trending at the moment.

The biggest strong points of this system are:

  • It offers precise and personalized recommendations.
  • It doesn’t have cold-start problems (poor performance due to lack of information).

The main drawback is:

  • It’s highly complex.

Machine Learning Techniques in Recommender Systems

It’s fair to say that machine learning is like the foundation stone of recommender systems. This sub-type of artificial intelligence (AI) represents the process of computers generating knowledge from data. We understand the “machine” part, but what does “learning” implicate? “Learning” means that machines improve their performance and enhance capabilities as they learn more information and become more “experienced.”

The four machine learning techniques recommender systems love are:

  • Supervised learning
  • Unsupervised learning
  • Reinforcement learning
  • Deep learning

Supervised Learning

In this case, algorithms feed off past data to predict the future. To do that, algorithms need to know what they’re looking for in the data and what the target is. The data in which we know the target label are named labeled datasets, and they teach algorithms how to classify data or make predictions.

Supervised learning has found its place in recommender systems because it helps understand patterns and offers valuable recommendations to users. It analyzes the users’ past behavior to predict their future. Plus, supervised learning can handle large amounts of data.

The most obvious drawback of supervised learning is that it requires human involvement, and training machines to make predictions is no walk in the park. There’s also the issue of result accuracy. Whether or not the results will be accurate largely depends on the input and target values.

Unsupervised Learning

With unsupervised learning, there’s no need to “train” machines on what to look for in datasets. Instead, the machines analyze the information to discover hidden patterns or similar features. In other words, you can sit back and relax while the algorithms do their magic. There’s no need to worry about inputs and target values, and that is one of the best things about unsupervised learning.

How does this machine learning technique fit into recommender systems? The main application is exploration. With unsupervised learning, you can discover trends and patterns you didn’t even know existed. It can discover surprising similarities and differences between users and their online behavior. Simply put, unsupervised learning can perfect your recommendation strategies and make them more precise and personal.

Reinforcement Learning

Reinforcement learning is another technique used in recommender systems. It functions like a reward-punishment system, where the machine has a goal that it needs to achieve through a series of steps. The machine will try a strategy, receive back, change the strategy as necessary, and try again until it reaches the goal and gets a reward.

The most basic example of reinforcement learning in recommender systems is movie recommendations. In this case, the “reward” would be the user giving a five-star rating to the recommended movie.

Deep Learning

Deep learning is one of the most advanced (and most fascinating) subcategories of AI. The main idea behind deep learning is building neural networks that mimic and function similarly to human brains. Machines that feature this technology can learn new information and draw their own conclusions without any human assistance.

Thanks to this, deep learning offers fine-tuned suggestions to users, enhances their satisfaction, and ultimately leads to higher profits for companies that use it.

Challenges and Future Trends in Recommender Systems

Although we may not realize it, recommender systems are the driving force of online purchases and content streaming. Without them, we wouldn’t be able to discover amazing TV shows, movies, songs, and products that make our lives better, simpler, and more enjoyable.

Without a doubt, the internet would look very different if it wasn’t for recommender systems. But as you may have noticed, what you see as recommended isn’t always what you want, need, or like. In fact, the recommendations can be so wrong that you may be shocked how the internet could misinterpret you like that. Recommender systems aren’t perfect (at least not yet), and they face different challenges that affect their performance:

  • Data sparsity and scalability – If users don’t leave a trace online (don’t review items), the machines don’t have enough data to analyze and make recommendations. Likewise, the datasets change and grow constantly, which can also represent an issue.
  • Cold start problem – When new users become a part of a system, they may not receive relevant recommendations because algorithms don’t “know” their preferences, past purchases, or ratings. The same goes for new items introduced to a system.
  • Privacy and security concerns – Privacy and security are always at the spotlight of recommender systems. The situation is a paradox. The more a system knows about you, the better recommendations you’ll get. At the same time, you may not be willing to let a system learn your personal information if you want to maintain your privacy. But then, you won’t enjoy great recommendations.
  • Incorporating contextual information – Besides “typical” information, other data can help make more precise and relevant recommendations. The problem is how to incorporate them.
  • Explainability and trust – Can a recommender system explain why it made a certain recommendation, and can you trust it?

Discover New Worlds with Recommender Systems

Recommender systems are growing smarter by the day, thanks to machine learning and technological advancements. The recommendations were introduced to allow us to save time and find exactly what we’re looking for in a jiff. At the same time, they let us experiment and try something different.

While recommender systems have come a long way, there’s still more than enough room for further development.

Related posts

Cyber Threat Landscape 2024: Human-Centric Cyber Threats
OPIT - Open Institute of Technology
OPIT - Open Institute of Technology
Apr 17, 2024 9 min read

Human-centric cyber threats have long posed a serious issue for organizations. After all, humans are often the weakest link in the cybersecurity chain. Unfortunately, when artificial intelligence came into the mix, it only made these threats even more dangerous.

So, what can be done about these cyber threats now?

That’s precisely what we asked Tom Vazdar, the chair of the Enterprise Cybersecurity Master’s program at the Open Institute of Technology (OPIT), and Venicia Solomons, aka the “Cyber Queen.”

They dedicated a significant portion of their “Cyber Threat Landscape 2024: Navigating New Risks” master class to AI-powered human-centric cyber threats. So, let’s see what these two experts have to say on the topic.

Human-Centric Cyber Threats 101

Before exploring how AI impacted human-centric cyber threats, let’s go back to the basics. What are human-centric cyber threats?

As you might conclude from the name, human-centric cyber threats are cybersecurity risks that exploit human behavior or vulnerabilities (e.g., fear). Even if you haven’t heard of the term “human-centric cyber threats,” you’ve probably heard of (or even experienced) the threats themselves.

The most common of these threats are phishing attacks, which rely on deceptive emails to trick users into revealing confidential information (or clicking on malicious links). The result? Stolen credentials, ransomware infections, and general IT chaos.

How Has AI Impacted Human-Centric Cyber Threats?

AI has infiltrated virtually every cybersecurity sector. Social engineering is no different.

As mentioned, AI has made human-centric cyber threats substantially more dangerous. How? By making them difficult to spot.

In Venicia’s words, AI has allowed “a more personalized and convincing social engineering attack.”

In terms of email phishing, malicious actors use AI to write “beautifully crafted emails,” as Tom puts it. These emails contain no grammatical errors and can mimic the sender’s writing style, making them appear more legitimate and harder to identify as fraudulent.

These highly targeted AI-powered phishing emails are no longer considered “regular” phishing attacks but spear phishing emails, which are significantly more likely to fool their targets.

Unfortunately, it doesn’t stop there.

As AI technology advances, its capabilities go far beyond crafting a simple email. Venicia warns that AI-powered voice technology can even create convincing voice messages or phone calls that sound exactly like a trusted individual, such as a colleague, supervisor, or even the CEO of the company. Obey the instructions from these phone calls, and you’ll likely put your organization in harm’s way.

How to Counter AI-Powered Human-Centric Cyber Threats

Given how advanced human-centric cyber threats have gotten, one logical question arises – how can organizations counter them? Luckily, there are several ways to do this. Some rely on technology to detect and mitigate threats. However, most of them strive to correct what caused the issue in the first place – human behavior.

Enhancing Email Security Measures

The first step in countering the most common human-centric cyber threats is a given for everyone, from individuals to organizations. You must enhance your email security measures.

Tom provides a brief overview of how you can do this.

No. 1 – you need a reliable filtering solution. For Gmail users, there’s already one such solution in place.

No. 2 – organizations should take full advantage of phishing filters. Before, only spam filters existed, so this is a major upgrade in email security.

And No. 3 – you should consider implementing DMARC (Domain-based Message Authentication, Reporting, and Conformance) to prevent email spoofing and phishing attacks.

Keeping Up With System Updates

Another “technical” move you can make to counter AI-powered human-centric cyber threats is to ensure all your systems are regularly updated. Fail to keep up with software updates and patches, and you’re looking at a strong possibility of facing zero-day attacks. Zero-day attacks are particularly dangerous because they exploit vulnerabilities that are unknown to the software vendor, making them difficult to defend against.

Top of Form

Nurturing a Culture of Skepticism

The key component of the human-centric cyber threats is, in fact, humans. That’s why they should also be the key component in countering these threats.

At an organizational level, numerous steps are needed to minimize the risks of employees falling for these threats. But it all starts with what Tom refers to as a “culture of skepticism.”

Employees should constantly be suspicious of any unsolicited emails, messages, or requests for sensitive information.

They should always ask themselves – who is sending this, and why are they doing so?

This is especially important if the correspondence comes from a seemingly trusted source. As Tom puts it, “Don’t click immediately on a link that somebody sent you because you are familiar with the name.” He labels this as the “Rule No. 1” of cybersecurity awareness.

Growing the Cybersecurity Culture

The ultra-specific culture of skepticism will help create a more security-conscious workforce. But it’s far from enough to make a fundamental change in how employees perceive (and respond to) threats. For that, you need a strong cybersecurity culture.

Tom links this culture to the corporate culture. The organization’s mission, vision, statement of purpose, and values that shape the corporate culture should also be applicable to cybersecurity. Of course, this isn’t something companies can do overnight. They must grow and nurture this culture if they are to see any meaningful results.

According to Tom, it will probably take at least 18 months before these results start to show.

During this time, organizations must work on strengthening the relationships between every department, focusing on the human resources and security sectors. These two sectors should be the ones to primarily grow the cybersecurity culture within the company, as they’re well versed in the two pillars of this culture – human behavior and cybersecurity.

However, this strong interdepartmental relationship is important for another reason.

As Tom puts it, “[As humans], we cannot do anything by ourselves. But as a collective, with the help within the organization, we can.”

Staying Educated

The world of AI and cybersecurity have one thing in common – they never sleep. The only way to keep up with these ever-evolving worlds is to stay educated.

The best practice would be to gain a solid base by completing a comprehensive program, such as OPIT’s Enterprise Cybersecurity Master’s program. Then, it’s all about continuously learning about new developments, trends, and threats in AI and cybersecurity.

Conducting Regular Training

For most people, it’s not enough to just explain how human-centric cyber threats work. They must see them in action. Especially since many people believe that phishing attacks won’t happen to them or, if they do, they simply won’t fall for them. Unfortunately, neither of these are true.

Approximately 3.4 billion phishing emails are sent each day, and millions of them successfully bypass all email authentication methods. With such high figures, developing critical thinking among the employees is the No. 1 priority. After all, humans are the first line of defense against cyber threats.

But humans must be properly trained to counter these cyber threats. This training includes the organization’s security department sending fake phishing emails to employees to test their vigilance. Venicia calls employees who fall for these emails “clickers” and adds that no one wants to be a clicker. So, they do everything in their power to avoid falling for similar attacks in the future.

However, the key to successful employee training in this area also involves avoiding sending similar fake emails. If the company keeps trying to trick the employees in the same way, they’ll likely become desensitized and less likely to take real threats seriously.

So, Tom proposes including gamification in the training. This way, the training can be more engaging and interactive, encouraging employees to actively participate and learn. Interestingly, AI can be a powerful ally here, helping create realistic scenarios and personalized learning experiences based on employee responses.

Following in the Competitors’ Footsteps

When it comes to cybersecurity, it’s crucial to be proactive rather than reactive. Even if an organization hasn’t had issues with cyberattacks, it doesn’t mean it will stay this way. So, the best course of action is to monitor what competitors are doing in this field.

However, organizations shouldn’t stop with their competitors. They should also study other real-world social engineering incidents that might give them valuable insights into the tactics used by the malicious actors.

Tom advises visiting the many open-source databases reporting on these incidents and using the data to build an internal educational program. This gives organizations a chance to learn from other people’s mistakes and potentially prevent those mistakes from happening within their ecosystem.

Stay Vigilant

It’s perfectly natural for humans to feel curiosity when it comes to new information, anxiety regarding urgent-looking emails, and trust when seeing a familiar name pop up on the screen. But in the world of cybersecurity, these basic human emotions can cause a lot of trouble. That is, at least, when humans act on them.

So, organizations must work on correcting human behaviors, not suppressing basic human emotions. By doing so, they can help employees develop a more critical mindset when interacting with digital communications. The result? A cyber-aware workforce that’s well-equipped to recognize and respond to phishing attacks and other cyber threats appropriately.

Read the article
Cyber Threat Landscape 2024: The AI Revolution in Cybersecurity
OPIT - Open Institute of Technology
OPIT - Open Institute of Technology
Apr 17, 2024 9 min read

There’s no doubt about it – artificial intelligence has revolutionized almost every aspect of modern life. Healthcare, finance, and manufacturing are just some of the sectors that have been virtually turned upside down by this powerful new force. Cybersecurity also ranks high on this list.

But as much as AI can benefit cybersecurity, it also presents new challenges. Or – to be more direct –new threats.

To understand just how serious these threats are, we’ve enlisted the help of two prominent figures in the cybersecurity world – Tom Vazdar and Venicia Solomons. Tom is the chair of the Master’s Degree in Enterprise Cybersecurity program at the Open Institute of Technology (OPIT). Venicia, better known as the “Cyber Queen,” runs a widely successful cybersecurity community looking to empower women to succeed in the industry.

Together, they held a master class titled “Cyber Threat Landscape 2024: Navigating New Risks.” In this article, you get the chance to hear all about the double-edged sword that is AI in cybersecurity.

How Can Organizations Benefit From Using AI in Cybersecurity?

As with any new invention, AI has primarily been developed to benefit people. In the case of AI, this mainly refers to enhancing efficiency, accuracy, and automation in tasks that would be challenging or impossible for people to perform alone.

However, as AI technology evolves, its potential for both positive and negative impacts becomes more apparent.

But just because the ugly side of AI has started to rear its head more dramatically, it doesn’t mean we should abandon the technology altogether. The key, according to Venicia, is in finding a balance. And according to Tom, this balance lies in treating AI the same way you would cybersecurity in general.

Keep reading to learn what this means.

Top of Form

Implement a Governance Framework

In cybersecurity, there is a governance framework called ISO/IEC 27000, whose goal is to provide a systematic approach to managing sensitive company information, ensuring it remains secure. A similar framework has recently been created for AI— ISO/IEC 42001.

Now, the trouble lies in the fact that many organizations “don’t even have cybersecurity, not to speak artificial intelligence,” as Tom puts it. But the truth is that they need both if they want to have a chance at managing the risks and complexities associated with AI technology, thus only reaping its benefits.

Implement an Oversight Mechanism

Fearing the risks of AI in cybersecurity, many organizations chose to forbid the usage of this technology outright within their operations. But by doing so, they also miss out on the significant benefits AI can offer in enhancing cybersecurity defenses.

So, an all-out ban on AI isn’t a solution. A well-thought-out oversight mechanism is.

According to Tom, this control framework should dictate how and when an organization uses cybersecurity and AI and when these two fields are to come in contact. It should also answer the questions of how an organization governs AI and ensures transparency.

With both of these frameworks (governance and oversight), it’s not enough to simply implement new mechanisms. Employees should also be educated and regularly trained to uphold the principles outlined in these frameworks.

Control the AI (Not the Other Way Around!)

When it comes to relying on AI, one principle should be every organization’s guiding light. Control the AI; don’t let the AI control you.

Of course, this includes controlling how the company’s employees use AI when interacting with client data, business secrets, and other sensitive information.

Now, the thing is – people don’t like to be controlled.

But without control, things can go off the rails pretty quickly.

Tom gives just one example of this. In 2022, an improperly trained (and controlled) chatbot gave an Air Canada customer inaccurate information and a non-existing discount. As a result, the customer bought a full-price ticket. A lawsuit ensued, and in 2024, the court ruled in the customer’s favor, ordering Air Canada to pay compensation.

This case alone illustrates one thing perfectly – you must have your AI systems under control. Tom hypothesizes that the system was probably affordable and easy to implement, but it eventually cost Air Canada dearly in terms of financial and reputational damage.

How Can Organizations Protect Themselves Against AI-Driven Cyberthreats?

With well-thought-out measures in place, organizations can reap the full benefits of AI in cybersecurity without worrying about the threats. But this doesn’t make the threats disappear. Even worse, these threats are only going to get better at outsmarting the organization’s defenses.

So, what can the organizations do about these threats?

Here’s what Tom and Venicia suggest.

Fight Fire With Fire

So, AI is potentially attacking your organization’s security systems? If so, use AI to defend them. Implement your own AI-enhanced threat detection systems.

But beware – this isn’t a one-and-done solution. Tom emphasizes the importance of staying current with the latest cybersecurity threats. More importantly – make sure your systems are up to date with them.

Also, never rely on a single control system. According to our experts, “layered security measures” are the way to go.

Never Stop Learning (and Training)

When it comes to AI in cybersecurity, continuous learning and training are of utmost importance – learning for your employees and training for the AI models. It’s the only way to ensure all system aspects function properly and your employees know how to use each and every one of them.

This approach should also alleviate one of the biggest concerns regarding an increasing AI implementation. Namely, employees fear that they will lose their jobs due to AI. But the truth is, the AI systems need them just as much as they need those systems.

As Tom puts it, “You need to train the AI system so it can protect you.”

That’s why studying to be a cybersecurity professional is a smart career move.

However, you’ll want to find a program that understands the importance of AI in cybersecurity and equips you to handle it properly. Get a master’s degree in Enterprise Security from OPIT, and that’s exactly what you’ll get.

Join the Bigger Fight

When it comes to cybersecurity, transparency is key. If organizations fail to report cybersecurity incidents promptly and accurately, they not only jeopardize their own security but also that of other organizations and individuals. Transparency builds trust and allows for collaboration in addressing cybersecurity threats collectively.

So, our experts urge you to engage in information sharing and collaborative efforts with other organizations, industry groups, and governmental bodies to stay ahead of threats.

How Has AI Impacted Data Protection and Privacy?

Among the challenges presented by AI, one stands out the most – the potential impact on data privacy and protection. Why? Because there’s a growing fear that personal data might be used to train large AI models.

That’s why European policymakers sprang into action and introduced the Artificial Intelligence Act in March 2024.

This regulation, implemented by the European Parliament, aims to protect fundamental rights, democracy, the rule of law, and environmental sustainability from high-risk AI. The act is akin to the well-known General Data Protection Regulation (GDPR) passed in 2016 but exclusively targets the use of AI. The good news for those fearful of AI’s potential negative impact is that every requirement imposed by this act is backed up with heavy penalties.

But how can organizations ensure customers, clients, and partners that their data is fully protected?

According to our experts, the answer is simple – transparency, transparency, and some more transparency!

Any employed AI system must be designed in a way that doesn’t jeopardize anyone’s privacy and freedom. However, it’s not enough to just design the system in such a way. You must also ensure all the stakeholders understand this design and the system’s operation. This includes providing clear information about the data being collected, how it’s being used, and the measures in place to protect it.

Beyond their immediate group of stakeholders, organizations also must ensure that their data isn’t manipulated or used against people. Tom gives an example of what must be avoided at all costs. Let’s say a client applies for a loan in a financial institution. Under no circumstances should that institution use AI to track the client’s personal data and use it against them, resulting in a loan ban. This hypothetical scenario is a clear violation of privacy and trust.

And according to Tom, “privacy is more important than ever.” The same goes for internal ethical standards organizations must develop.

Keeping Up With Cybersecurity

Like most revolutions, AI has come in fast and left many people (and organizations) scrambling to keep up. However, those who recognize that AI isn’t going anywhere have taken steps to embrace it and fully benefit from it. They see AI for what it truly is – a fundamental shift in how we approach technology and cybersecurity.

Those individuals have also chosen to advance their knowledge in the field by completing highly specialized and comprehensive programs like OPIT’s Enterprise Cybersecurity Master’s program. Coincidentally, this is also the program where you get to hear more valuable insights from Tom Vazdar, as he has essentially developed this course.

Read the article