Artificial Intelligence

Experts predict how close AI is to passing the hardest test in the world

Can you provide a translation of an ancient Palmyrene script found on a Roman tombstone? How many paired tendons are supported by a particular sesamoid bone in a hummingbird?

These are just two of the many varied and challenging questions submitted to Humanity’s Last Exam(opens in new window), or HLE, the apparently unsolvable test reserved only for the best and brightest. But it’s not meant for us.

Ultimate academic test for AI

Don’t let the apocalyptic name fool you. HLE isn’t about humans becoming irrelevant. It’s about celebrating what we know that AI can’t even touch yet. It was created to determine if AI models such as ChatGPT and Gemini can answer the most difficult questions experts could come up with.

Basically, HLE was specifically designed to see exactly where today’s AI fails and what remains out of reach for existing AI technology. This new benchmark was introduced in a study published in the journal ‘Nature’(opens in new window).

HLE is a truly collaborative effort, with about 2 500 questions from close to 1 000 contributors affiliated with over 500 institutions across 50 countries. Contributors are mainly experts in the sciences, humanities and arts, covering more than 100 highly specialised fields.

“What made this project extraordinary was the scale,” commented Tung Nguyen, associate professor in the Department of Computer Science and Engineering at Texas A&M, in a news release(opens in new window). “That diversity is exactly what exposes the gaps in today’s AI systems — perhaps ironically, it’s humans working together.”

AI systems didn’t exactly ace this tough test – at least not early on. Initial results in 2025 showed that many of the AIs scored less than 10 % on the exam. However, in March of this year, Gemini 3.1 Pro achieved 45.9 % accuracy, followed closely behind by GPT-5.4 at 40.3 %.

Surpassing the boundaries of human knowledge

“When AI systems start performing extremely well on human benchmarks, it’s tempting to think they’re approaching human‑level understanding,” explained Nguyen. “But HLE reminds us that intelligence isn’t just about pattern recognition — it’s about depth, context and specialized expertise.”

Nguyen contributed 73 of the questions – the second-highest number – and wrote the most questions in math and computer science. “For now, Humanity’s Last Exam stands as one of the clearest assessments of the gap between AI and human intelligence, and despite rapid technological advances, it remains wide.”

Nguyen emphasised that when AI surpasses traditional metrics, the resulting gap creates challenges that are more than just academic. “Without accurate assessment tools, policymakers, developers and users risk misinterpreting what AI systems can actually do. Benchmarks provide the foundation for measuring progress and identifying risks.”

HLE is a reality check for AI, proving that our unique knowledge still sets the bar higher than any algorithm can reach. “This isn’t a race against AI,” Nguyen concluded. “It’s a method for understanding where these systems are strong and where they struggle. That understanding helps us build safer, more reliable technologies. And, importantly, it reminds us why human expertise still matters.”

Satellite and citizen data feed scalable biodiversity indicators, helping planners, consultants and managers compare risks, target surveys and act sooner

Biodiversity decisions often depend on incomplete data. A site might have a few surveys conducted in different years by different teams using different methods. This makes it difficult to compare locations, spot trends early and explain trade-offs when new infrastructure, land use or conservation measures are being considered.

The EU-funded GUARDEN project(opens in new window) set out to improve the situation with a practical monitoring toolbox. By combining satellite-based indicators, AI modelling, citizen-science platforms, acoustic sensors and augmented reality tools, it aims to enable more frequent updates to biodiversity and ecosystem services information and to support the application of that information in real-world planning.

Satellite and AI indicators that scale across landscapes

GUARDEN’s most consistent results came from combining Earth observation with AI-based ecological modelling. As project coordinator Pierre Bonnet explains, “In real-world conditions, the combination of Earth observation data with AI-based ecological modelling proved to be one of the most robust and scalable monitoring approaches within GUARDEN.” Those satellite-derived habitat indicators were designed to be repeatable across different landscapes and governance contexts, supporting comparisons between areas and scenarios.

To make the outputs easier to use, GUARDEN made its map set accessible through the GeoPl@ntNet web mapping tool(opens in new window). The approach works best for landscape-scale patterns and trend tracking, while fine-scale local features and rare species can still be hard to capture without targeted fieldwork.

Blending methods to reduce blind spots

GUARDEN treated monitoring as a combination problem, because each method breaks in a different way. Citizen science observations via specific platforms, namely MINKA and Pl@ntNet, can provide valuable field records; however, they are sensitive to uneven sampling effort and observer bias. Acoustic sensors can be highly repeatable, but performance drops when recordings are masked by wind, rain and machinery or when target species are rare.

The project’s takeaway is blunt: “One key lesson was that no single monitoring method is sufficient on its own. The most reliable results emerged when tools were explicitly combined, with each one compensating for the blind spots of the others.” Before data entered models, GUARDEN added quality-control layers including duplicate and anomaly checks, confidence scoring for identifications, uncertainty thresholds, and cross-validation using reference datasets and expert-reviewed samples.

Ground validation, benchmarking and what changed on the ground

Validation was iterative, mixing targeted field surveys, existing monitoring programmes, expert assessments and citizen observations. A notable technical step was benchmarking GUARDEN’s deep learning framework through the GeoLifeCLEF(opens in new window) and PlantCLEF(opens in new window) international challenges, providing external performance benchmarks and helping calibrate models.

When the maps and field observations diverged, the response was to diagnose and adjust rather than discard. Bonnet notes, “When discrepancies emerged between field observations and model outputs, they were treated as signals for improvement rather than failures.” Causes included training data gaps, scale effects and local management factors, with local refinements made where needed, including for legal expectations.

In practice, GUARDEN outputs informed route option reviews for transport infrastructure, helped concentrate field surveys on model-flagged hotspots, supported management prioritisation in peri-urban habitats and enabled quicker flagging of potentially invasive species using citizen data.

Will AI really replace teachers? Does it stifle creativity? Is more AI better? Stanford experts weigh in on common assumptions

When you think about AI in classrooms, what emotions come up? Skepticism? Amazement? Curiosity? Fear?

As AI increasingly infuses every aspect of our lives, educators, students, parents, policymakers, and researchers are returning to fundamental questions about learning and school.

In the winter of 2025-2026, the Stanford Accelerator for Learning hosted five events convening leaders to discuss how AI can support learning, how to develop effective AI tools, how to protect young people from AI’s risks, and what should remain fundamentally human. The conversations challenged five common myths about AI and education.

1. Myth: AI will replace teachers.

Reality: In an increasingly tech-centered world, human connection is indispensable.

One of the most persistent fears about AI and education is that technology will eventually replace teachers. But speakers said the opposite: in an increasingly digital world, the classroom is more vital than ever as a space for human relationships and peer communities.

“We know that learning is fundamentally cultural and social,” said Daniela Di Giacomo, associate professor at the University of Kentucky, at the fourth annual AI+Education Summit, co-hosted with the Stanford Institute for Human-Centered Artificial Intelligence (HAI). “[When] we think about how people learn in a digital age and hyperpartisan age, that should actually amplify the need for better pedagogical instruction.”

Teacher Mike Taubman, who serves as AI innovation lead at Uncommon Schools, agreed. “The classroom is taking on an almost sacred dimension for me now, where people are gathering together to be young and human together, and grow up together and learn to argue in a very complicated country together.”

Teachers also play a key role in students getting value from AI tools when they are used. At the Youth-Powered AI Day of Learning, a gathering of teachers and middle and high schoolers hosted by the Accelerator’s Equity in Learning initiative, Nathan Pierce, a teacher at Design Tech High School, suggested the idea that teachers may eventually play more of a coaching role.

“Getting an adult to work with [a learner] can get them to work on that platform better … the adult needs to be there to motivate and engage them,” said Susanna Loeb, faculty director of the Accelerator’s SCALE initiative.

2. Myth: AI makes it too easy to cheat.

Reality: We need to rethink assessment.

At the Accelerator’s conference on Responsible Assessment in the AI Era, in collaboration with ETS, speakers reframed cheating concerns as an opportunity to fundamentally rethink what and how we assess.

“We have to move beyond just thinking about a test as something we do at the end,” said Amit Sevak, CEO of ETS, one of the world’s largest educational testing organizations, which designs and administers the TOEFL, Praxis, and GRE. Emerging approaches include regular, data-driven formative assessments; scenario-based design that tests adaptability; and AI-powered personalization of tests. Best practices for any assessment that involves AI include co-design with educators and continuous human oversight.

“I think the days of us testing on sheer rote knowledge in a homework question is probably over. But these more delicate questions about synthesis, analysis, and integrating, that you can answer in a layered, multi-stage framework – these are the things that [AI] tools are effectively useless for,” said Paul Nuyujukian, assistant professor of neuroscience at Stanford, at the Accelerator’s third annual Accelerate Edtech Impact Summit. “The beautiful thing is that that’s the intuition you actually want to impart in your classroom.”

Dan Schwartz, dean of the Stanford Graduate School of Education (GSE) and faculty director of the Accelerator, agreed. “Our knowledge bases, our tools, our circumstances, our jobs, continue to change, so we need instruction that produces adaptive learners and assessments that can tell.”

3. Myth: AI stifles creativity.

Reality: With intentional design, AI can be a tool to spark creativity, but we need to be careful about potential drawbacks.

At the AI+Education Summit, Mehran Sahami, chair of Stanford’s computer science department, challenged the assumption that AI kills creativity. “A lot of critics say AI [is] just trained on certain information and [it] can only regurgitate that information … what do we do for most of education? Exactly the same thing. And somehow we expect students to be able to generate novel results.” The question, he argued, is finding ways that AI can help humans produce novel ideas, and crafting learning experiences that focus on the process of creation rather than the product.

Hari Subramonyam, faculty affiliate of the Accelerator, showed how AI can expand creative possibilities for students. “AI can provide the infrastructure to lower the floor for creation to get students started,” he said, showcasing examples like animation tools that allow hands-on exploration of physics concepts. It can also “raise the ceiling” of what is technically feasible, allowing learners to focus on higher-order thinking. “Creation helps learners organize and structure knowledge in more meaningful, usable ways, and AI should support this,” he said.

However, speakers cautioned that using AI for creative tasks can backfire. Accelerator Faculty Affiliate Guilherme Lichand shared research from middle schools in Brazil showing that while AI assistance helped students on immediate creative tasks, it led to significantly worse performance on subsequent tasks when the AI was removed. “You start thinking that the AI is more creative than you,” he explained, underscoring the importance of intentional design.

4. Myth: The more new AI tools for learning, the better.

Reality: It’s easy to make an AI tool, but harder to make one backed by science or research.

“We’ve democratized the ability to create products,” said Accelerator Faculty Affiliate Susan Athey, a professor at Stanford Graduate School of Business and an advisor to the World Bank. “People who have ideas, even non-technical people … now can make their ideas a reality. That’s very exciting. But what that’s starting to look like on the ground, is now that we have too many pilots, and still not enough implementations that are actually effective.”

The solution, speakers emphasized, is grounding AI development in learning science and iterative design. Loeb urged developers to start with research. “The first step is to take what we know and apply it, and don’t make the obvious mistakes that we know from everything we’ve done in edtech,” she said. Then, programs should collect data on implementation and engagement through randomization and experimentation. “If you’re rolling out a program to 10,000 schools with 100,000 students, do it in a way where you learn something,” she said.

James Landay, co-founder and co-director of Stanford HAI and a faculty affiliate of the Accelerator, argued, “We need to go beyond user-centered design to human-centered AI design,” bringing students, teachers, families, and learning experts into the process and considering societal-level effects, particularly when it comes to tools that scale widely.

The Accelerator’s Create+AI Challenge brought 10 cross-sector teams to campus to pitch projects that put educators and learners at the heart of AI design. The projects, which aimed either to augment teaching, augment learning, or augment career opportunities, were judged not on how impressive the technology, but rather on their basis in research and design principles, and potential impact on learning.

Equally important is recognizing when an AI tool isn’t necessary at all. “If you can do it with a paper and pencil, or in-person – just give a kid a hug or have a chat – just do that! Use the technology for things that are really transformational,” said Rebecca Winthrop, senior fellow at the Brookings Institution.

5. Myth: Banning AI from classrooms is the best way to protect students from its risks.

Reality: AI is here to stay. Integrating it responsibly – with ethical guidelines and AI literacy – will prepare students to thrive.

Seventy-two percent of K-12 students routinely use generative AI, but only 28% can accurately describe how it works, cited Ronit Levavi Morad, senior director at Google Research, at the AI+Education Summit. How do we ensure the safety of young people with tools that they may not understand, are not necessarily designed for them, and are often used outside of school contexts? “What’s the balance between ‘protect our kids’ and ‘prepare our kids’? That is a real tension lived on a daily basis in classrooms and in homes,” said Winthrop.

Speakers across events and panels referenced lessons learned from prior technologies like social media, which schools initially ignored or banned. “This is just the most recent iteration of the thing that will be in their life, and we either teach them how to use it, or they will be used by it,” said Kirsten Baesler, U.S. assistant secretary for elementary and secondary education, at the Accelerate Edtech Impact Summit.

The path forward requires both safety measures and AI literacy education. Erin Mote, CEO at InnovateEDU, argued that “safety is not the counterpolarity of innovation.” Rather, through a focus on safety, “we can actually unlock the types of positive use cases in AI, in schools and education, that move us beyond fear and towards knowledge, and the types of learning experiences we want for young people.”

Taubman’s “AI driver’s license” program exemplifies this approach. “The idea is to map that quintessential adolescent experience of getting your driver’s license onto this AI moment,” he said. “The whole idea, as you can imagine, is to get students into the driver’s seat and not the passenger seat when it comes to AI.” The impact of the program: “They start to realize that their voices matter right now and they can start to take part in shaping this world that they’re graduating into.”

Similarly, the AI Quests game, co-designed by Stanford scholars and a team from Google Research, aims to move students from passive users of AI to critical thinkers, said Victor Lee, faculty lead for AI+Education at the Accelerator and a co-creator of the game. “We’re positioning students and educators to have a sense of agency in how we use AI, how we create AI, and how we evaluate AI.”

Image: Allison Shelley / EDUimages

World’s largest creativity experiment tested AI against humans

We’ve always viewed creativity as a unique by-product of human consciousness, as a defining human trait. This is what separates us from the tech world and machines.

However, recent research has shown that ChatGPT can already outdo the average person on creative tasks. Can machines match or even surpass human creativity? Are creative individuals such as writers, designers and artists going to be replaced?

AI versus human creativity

A research team led by Karim Jerbi from the Department of Psychology at the University of Montreal in Canada carried out the biggest comparison between human and machine creativity up to now. Using a standard creativity test, they pitted over 100 000 participants from Australia, Canada, New Zealand, the United Kingdom and the United States against nine of the most advanced AI systems.

The results revealed that GPT-4 scored higher than the average human on the test. Google’s Gemini Pro matched average human performance.

However, GPT-4 kept using the same words again and again. ‘Microscope’ appeared in 70 % of its responses, followed by ‘elephant’ (60 %). Because it was natural for the volunteers to avoid repeating themselves, the most common word was ‘car’ (1.4 %), followed by ‘dog’ (1.2 %) and then ‘tree’ (1 %).

“Even though AI can now reach human-level creativity on certain tests, we need to move beyond this misleading sense of competition,” explained study co-author Jerbi in a news release(opens in new window). “Generative AI has above all become an extremely powerful tool in the service of human creativity: it will not replace creators, but profoundly transform how they imagine, explore, and create — for those who choose to use it.”

As part of the test, the volunteers were asked to name 10 words in under four minutes that are as unrelated to each other as possible. An individual scored higher on creativity when the meanings of the words were further apart.

Creative talent triumphs

On the other hand, the top 10 % of creative people consistently outperformed every tested AI system. The findings were published in the journal ‘Scientific Reports’(opens in new window).

The researchers also tested AI models and humans on creative writing tasks, including haikus, film summaries and short stories. Here too, the most creative individuals outperformed AI, especially in poetry and plot synopses.

An interesting outcome was that GPT-4 Turbo, a faster, cheaper and more improved version of GPT-4, performed much worse on the test. This means that newer AI models aren’t necessarily more creative. The researchers suggest that the reason for this is newer releases are engineered for greater speed and cost-effectiveness, possibly at the expense of creativity.

“Our study shows that some AI systems based on large language models can now outperform average human creativity on well-defined tasks,” commented Jerbi. “This result may be surprising — even unsettling — but our study also highlights an equally important observation: even the best AI systems still fall short of the levels reached by the most creative humans.”

“By directly confronting human and machine capabilities, studies like ours push us to rethink what we mean by creativity,” he concluded.

Image: © www.thesmartcityjournal.com

Europe’s AI researchers are joining forces through an EU-funded digital platform that allows them to share tools, data and computing power to drive collaboration and innovation

Across Europe, researchers are using AI to tackle everything from underwater noise pollution to media fact-checking and smarter farming. Until recently, there was no widely used common gateway where they could easily share tools, data and computing power, but that is all changing. Europe’s AI research and innovation is now being brought together on a shared digital platform called AI-on-Demand (AIoD) – an EU-funded online hub designed to help researchers, startups, companies and public authorities collaborate and experiment more easily.

A single hub for European AI

Tanvir Singh Badwal from University College Cork in Ireland has been helping guide the development of AIoD since 2022. “In a nutshell, AIoD is a place where researchers and industry can access resources, use services and even develop new ones,” he said. The platform did not appear overnight. It began in 2019 under the EU-funded AI4EU project, coordinated by Thales in France, which first laid the foundations for a shared European AI hub. It was then significantly expanded by AI4EUROPE, a consortium of 24 institutions from 15 countries led by Badwal and University College Cork between 2022 and 2025. Today, the next phase is being driven by the DeployAI collaboration, funded by the Digital Europe Programme, which is working to bring the platform to market, expanding its use to industry and the public sector. The need for such a hub is clear. While the US and China dominate AI development with large tech giants and powerful platforms, Europe’s AI landscape is more fragmented. It consists of smaller players operating under different national rules, funding schemes and data standards. AIoD was created to help connect these dots. Rather than replacing existing initiatives such as Gaia-X or the European Open Science Cloud, the platform complements them by acting as a practical entry point for AI tools and collaboration. By making it easier to discover datasets, algorithms, computing resources and partners, the platform aims to speed up innovation and help European AI solutions move more quickly from research labs into real-world use. This all fits into the EU’s wider AI strategy to strengthen Europe’s research and industry, while keeping AI aligned with democratic values, fundamental rights and the rule of law.

From platform to practice

What does that look like in real life? Imagine a small agritech startup developing a tool to help farmers optimise irrigation and fertiliser use. The team could use AIoD to search for satellite and soil datasets and build a prototype predictive model using the platform’s AI Builder. This online tool allows users to create AI workflows through a visual interface without any need for heavy coding. They could then move to the Research and Innovation AI Lab (RAIL) – an online environment where experiments can be run directly on the platform. Through RAIL, users can access integrated high-performance computing resources to test and refine their models at scale. Once the tool is ready, the startup can use the platform’s community features to connect with mentors, find collaborators in other countries and share its solution across Europe. As a result, farmers could then receive forecasts days in advance and reduce water and fertiliser use.

One-stop shop for AI resources

AIoD combines several functions in one place: part search engine, part marketplace and part online laboratory. Users can browse datasets, scientific publications, educational materials, software components and pre-trained models. They can also access advanced computing power via links to European supercomputing centres such as the Barcelona Supercomputing Center and the LUMI system in Finland. Importantly, the platform does not host all these resources itself. Instead, it aggregates material from other established AI and open-source platforms, including Hugging Face, Bonseyes and OpenML. “Whatever has been uploaded onto these other platforms is fetched into ours,” Badwal explained. “The idea is that researchers can search in one place and access a much broader range of resources.” This matters because high-quality data and powerful computing resources are often concentrated in well-funded corporations or elite institutions. By lowering access barriers, AIoD aims to level the playing field for smaller labs and startups.

Keeping tools alive beyond projects

Another goal is to ensure that tools developed in EU research projects do not disappear once funding ends. One example is the Responsible Robotics Compass, or RoboCompass, developed within an earlier EU-funded project and now available through AIoD. This self-assessment tool looks at non-technical aspects of responsible robotics. It is designed to help researchers assess how well their robots align with public expectations and concerns. “You evaluate your robot according to various socio-economic, environmental, ethical and legal considerations,” said Joana Martinheira, a consultant at Portuguese communication agency LOBA, who was involved in the technical development of the online platform. “It’s like a quiz. At the end, you receive a score and recommendations on how to improve.” By hosting such tools, AIoD gives research outputs a longer life and wider audience.

Building an AI ecosystem

For those looking to build networks, the platform also includes mentoring sections, discussion forums and directories of AI projects and facilities across Europe. “AIoD simplifies collaborations and synergies between all the players in the ecosystem,” Martinheira said. By turning scattered tools and isolated projects into a shared, searchable and usable resource, AIoD aims to strengthen collaboration and sharpen Europe’s edge in AI. With DeployAI now focused on scaling the platform and linking it more closely to industry and public services, the ambition is clear: to give Europe not just strong AI research, but a functioning AI ecosystem capable of turning ideas into practical solutions for businesses, governments and everyday life.

Text : Michael Allen

This article was originally published in Horizon the EU Research and Innovation Magazine.

AI  enhances protection but does not replace human expertise. Over 4.7 Million Cybersecurity Vacancies Worldwide

Fortinet, the global cybersecurity leader driving the convergence of networking and security, today released its 2025 Global Cybersecurity Skills Gap Report, shedding light on the new and persistent challenges organizations face due to the cybersecurity skills gap. The global survey’s key findings include:

• As organizations are turning increasingly to AI to strengthen their security postures and fill gaps, they also acknowledge that AI may be used against them as an engine of new or improved cyberattacks, especially given the lack of AI skillsets across teams.
• Lack of cybersecurity awareness and training remains the top cause of breaches.
• Boards lack cyber knowledge, despite it being a priority.
• Organizations want cybersecurity personnel with certifications.

In Saudi Arabia, digital transformation under Vision 2030 has accelerated cloud adoption, e-government services, and smart city development, making AI-driven cybersecurity capabilities critical. The Kingdom’s National Cybersecurity Authority (NCA) emphasizes local talent development. However, demand still far outpaces supply.

“This year’s survey further underscores the urgent need to invest in cybersecurity talent,” said Carl Windsor, CISO at Fortinet. “Without closing the skills gap, organizations will continue to face rising breach rates and escalating costs. The findings highlight an inflection point for both public and private sectors: without bold action to build and retain cybersecurity expertise, the risks and costs will only continue to grow for our society.” 

“As Saudi Arabia accelerates its digital transformation under Vision 2030, cybersecurity has become a foundational pillar for national resilience and economic growth. The Fortinet Skills Gap Report confirms the on the ground observation that while AI is essential to enhancing cyber defense, the Kingdom must prioritize upskilling the cybersecurity workforce to fully realise its benefits. Strategic initiatives, such as the Saudi Cybersecurity Workforce Development Program and increased access to industry recognised certifications, are vital to building local capabilities. Public-private partnerships will be key to equipping local talent with the AI expertise needed to protect critical infrastructure, government services, and the broader digital economy.” said Sami Alshwairakh, Senior Regional Director Sales at Fortinet

Report Links Cyber Skills Gap to Escalating Security and Financial Risks

As cyberthreats continue to escalate, organizations face the reality that security attacks are not just a possibility but a certainty. At the same time, an estimated global shortfall of more than 4.7 million skilled professionals leads to critical security roles being unfilled at a time when they are needed most. Key findings about the impact of the skills gap on organizations globally include: 

• The volume of breaches organizations experience is increasing year over year. According to the 2025 Fortinet Global Skills Gap Report, 86% of organizations experienced at least one cyber breach in 2024, with nearly one-third (28%) reporting five or more. These figures mark a significant increase from 2021, when the inaugural Fortinet Global Skills Gap Report was released, in which 80% of organizations reported breaches, and only 19% faced five or more. 
• The cybersecurity skills shortage is a key contributor to increased breaches. More than 50% of those surveyed (54%) indicated a lack of IT security skills and training as one of the leading causes of breaches in their organizations. 
• Financial ramifications of breaches remain significant. More than half (52%) of surveyed organizations say cyber incidents cost them over $1 million in 2024, consistent with the prior year’s findings and sharply up from 38% in 2021.

In the Kingdom, government and financial services sectors are high-value targets. Saudi organizations report a growing number of incidents, especially around identity fraud, phishing, and AI-driven attacks. Regulatory frameworks like the Essential Cybersecurity Controls (ECC) mandate stronger cyber resilience, but local talent development still struggles to keep pace. 

AI Could Ease Strain on Security Teams, but Lack of Expertise Is a Growing Risk

While AI offers critical relief amid ongoing cyber skills shortages, organizations may not yet be fully prepared to harness its potential securely. This year’s survey found:

• Security technology with AI capabilities has been widely adopted. An overwhelming 97% of organizations surveyed are either already using or plan to implement AI-enabled cybersecurity solutions, with threat detection and prevention cited as the top areas of interest for applying AI in cybersecurity. 
• AI can help alleviate the burden on short-staffed security teams. 87% of cybersecurity professionals expect AI to enhance their roles, rather than replace them, offering efficiency and relief amid skills shortages.  
• While AI can help security teams, teams lack AI skillsets to unlock the technology’s full potential. A majority of those surveyed (80%) say AI is helping their IT and security teams become more effective, but nearly half (48%) of IT decision makers point to a lack of staff with sufficient AI expertise as the biggest challenge to successful implementation. Seventy-six percent of organizations that suffered nine or more cyberattacks in 2024 had AI tools in place, suggesting that adoption alone isn't enough without the right expertise.

Saudi enterprises are early adopters of AI in cybersecurity, especially in government, oil & gas, and banking sectors. Yet, access to AI-specialized cybersecurity professionals remains limited, prompting initiatives like the Saudi Cybersecurity Workforce Development Program to address this gap. 

As Board-Level Focus on Cybersecurity Grows, Understanding of AI Impact Lags 

When it comes to the board of directors’ understanding of cybersecurity’s role at their organization, the report revealed the following: 

• Cybersecurity prioritization at the board level is on the rise with 76% of boards increasing their focus on the issue in 2024. Nearly all organizations now view cybersecurity as both a business (96%) and financial (95%) priority. 
• Board members aren’t as aware of the potential risks that AI use poses to their organizations. Fewer than half (49%) of all respondents indicated their boards fully understand the risks posed by AI, with awareness closely linked to whether their organizations are already deploying AI in their cybersecurity programs. 

Upskilling Remains a Focus in Addressing the Skill Gap

As the cyber skills shortage persists, other key findings from the report include:

• Certifications continue to be highly valued by employers. Eighty-nine percent of IT decision-makers prefer to hire candidates who hold certifications. Most respondents said certifications validate cybersecurity knowledge (67%), demonstrate an ability to stay current in a fast-evolving field (61%), and indicate familiarity with key vendor tools (56%). 
• Organizational support for funding certifications has declined. Only 73% of respondents now say they are willing to pay for employees to obtain certifications, down from 89% in 2023. 

National initiatives like the Saudi Digital Academy and partnerships with global firms are working to close the cyber talent gap by subsidizing professional certifications and upskilling programs. 

Closing the Skills Gap Is Critical to Business Resilience

The 2025 Cybersecurity Skills Gap Report makes clear that cybersecurity has become a board-level priority, driven by the rise of AI and the escalating risks to business operations. Closing the global skills gap remains essential. Organizations must rethink hiring practices, tap into underutilized talent pools, and invest in training and upskilling to build and retain the expertise they need. This requires a coordinated approach grounded in three key pillars: raising awareness and education, expanding access to targeted training and certification, and embracing advanced security technologies. 

To help organizations address the challenges they face as a result of the cyber skills gap, the award-winning Fortinet Training Institute, one of the industry's broadest training and certification programs, is dedicated to making cybersecurity certification and new career opportunities available to all populations, including a Security Awareness Training service for organizations to develop a cyber-aware workforce. The Security Awareness and Training service offers AI-focused modules to enhance understanding of AI and the role it plays in cybersecurity, including an introduction to GenAI and curriculum around AI-powered threats, covering the various methods that cybercriminals use when harnessing AI to create and enhance cyberattacks. 

Additionally, as part of Fortinet’s commitment to addressing this growing challenge, Fortinet is on track to train 1 million people in cybersecurity around the world by the end of 2026, since setting that pledge in 2021.

About the Fortinet Skills Gap Survey

• The survey was conducted among over 1,850 IT and cybersecurity decision-makers from 29 different countries and locations.
• Survey respondents come from a range of industries, including technology (22%), manufacturing (16%), and financial services (12%).

Additional Resources

• Learn more about the Fortinet Training Institute programs, helping address the cyber skills gap through upskilling and reskilling, and free training. 
• Visit fortinet.com/trust to learn about Fortinet innovation, collaboration partners, product security processes, and enterprise-grade products. 
• Read about how Fortinet customers are securing their organizations.
• Learn about Fortinet's commitment to product security and integrity, including its responsible product development and vulnerability disclosure approach and policies. 
• Follow Fortinet on X, LinkedIn, Facebook, and Instagram. Subscribe to Fortinet on our blog or YouTube.