As artificial intelligence accelerates the discovery of software vulnerabilities, organisations are facing an increasing volume of cybersecurity alerts. But identifying a vulnerability, establishing whether it affects your website, and deciding what action to take are three very different things.
TL;DR – The key takeaways
- Vulnerability reports are increasing rapidly. AI-assisted security research is accelerating the discovery of weaknesses in software and infrastructure.
- Not every vulnerability has a CVE. CVE databases are important security resources, but they don’t capture every vulnerability or provide a complete picture of cybersecurity.
- Not every CVE affects your website. A reported vulnerability might concern software you don’t use, functionality you haven’t enabled, or a configuration that doesn’t apply to your environment.
- Not all vulnerabilities carry the same risk. Severity, exploitability, exposure and business impact all influence how urgently action is required.
- Patching isn’t always possible or sufficient. Alternative protections, configuration changes and additional investigation may be necessary.
- Security scanners aren’t infallible. They can report vulnerabilities that have already been addressed or aren’t exploitable in a particular environment.
- Good cybersecurity is continuous risk management, not simply a checklist of installed patches.
What exactly is a CVE?
CVE stands for Common Vulnerabilities and Exposures. It’s an internationally recognised system for identifying and cataloguing publicly disclosed cybersecurity vulnerabilities.
Established in 1999, the CVE programme assigns unique identifiers, such as CVE-2026-12345, allowing software developers, security researchers, hosting providers and customers to refer to the same vulnerability consistently.
The CVE Program maintains this identification system. Other organisations, including the US National Institute of Standards and Technology (NIST), provide additional analysis through resources such as the National Vulnerability Database (NVD).
Importantly, a CVE is an identifier, not a diagnosis of your website’s security.
A published CVE indicates that a particular vulnerability has been identified. It doesn’t automatically mean your website uses the affected software, is exposed to the vulnerability, or has been compromised.
Understanding that distinction is fundamental to effective cybersecurity.
CVEs aren’t the whole cybersecurity picture
Although CVEs are enormously useful, they’re neither an exhaustive catalogue of security vulnerabilities nor a complete measure of cybersecurity.
Not every vulnerability receives a CVE identifier.
Some vulnerabilities remain undiscovered or undisclosed. Others arise from insecure configurations, inadequate access controls, application-specific weaknesses or circumstances that may never result in a published CVE.
Furthermore, vulnerability information isn’t confined to the CVE programme.
Other important resources include:
- NIST National Vulnerability Database – provides additional analysis of published CVEs, including severity assessments and affected software.
- GitHub Advisory Database – tracks vulnerabilities affecting open-source software, including advisories that don’t have CVE identifiers.
- OSV.dev – provides vulnerability information for open-source software packages, including precise affected-version information.
- CISA Known Exploited Vulnerabilities Catalogue – identifies CVEs with evidence of active exploitation.
- Software vendor advisories – provide security information, patches and configuration guidance directly from software developers and maintainers.
These sources overlap, but they serve different purposes. Some identify vulnerabilities, others provide additional technical detail, and others help security teams determine which threats deserve immediate attention.
The absence of known CVEs doesn’t mean a website or server is secure.
Effective cybersecurity requires looking beyond any single database to consider the broader security of software, infrastructure, configurations and operational practices.
The AI effect: More vulnerabilities, discovered faster
The volume of publicly reported software vulnerabilities has increased dramatically.
Approximately 40,000 CVE records were published in 2024, compared with more than 48,000 in 2025.
Artificial intelligence is increasingly influencing this landscape.
AI-assisted tools can analyse enormous quantities of source code, identify potential weaknesses and help security researchers investigate vulnerabilities more efficiently.
Unfortunately, malicious actors can use similar capabilities to discover weaknesses and assist in developing attacks.
However, AI isn’t the only reason vulnerability reporting is increasing. Greater security research activity, improved disclosure practices and increasingly automated discovery processes are also contributing.
And importantly, more reported vulnerabilities don’t necessarily mean every website has become proportionately less secure.
What they do mean is that separating genuine risks from background noise is becoming increasingly important.
Why isn’t patching straightforward?
When a vulnerability is reported, it’s understandable for a website owner to ask:
“Are we affected, and has it been patched?”
Those are entirely reasonable questions. Unfortunately, the answers aren’t always straightforward.
Here are some of the considerations involved.
1. The vulnerability might not affect your website
A CVE may concern a particular software package, version, module or configuration.
For example, a vulnerability affecting an optional Apache web server module may be irrelevant if that module isn’t installed or enabled.
A server could therefore be running Apache while remaining unaffected by a particular Apache-related CVE.
Establishing whether a vulnerability actually applies is the first step. Not every published CVE requires action on every server.
2. Vulnerabilities have different levels of risk
Some vulnerabilities could allow attackers to take control of a server or access sensitive information. Others may have limited consequences or require very specific circumstances to exploit.
Security professionals consider several factors:
- The technical severity of the vulnerability.
- Whether attackers are actively exploiting it.
- Whether the affected system is publicly accessible.
- Whether existing security protections reduce exposure.
- The potential consequences for the organisation.
Industry tools such as CVSS (Common Vulnerability Scoring System) help assess severity, while EPSS (Exploit Prediction Scoring System) estimates the likelihood of exploitation.
These are useful inputs, but neither replaces an assessment of the actual environment.
The goal is to prioritise genuine risk, not treat every vulnerability as equally urgent.
3. Sometimes a patch doesn’t exist yet
Discovering a vulnerability doesn’t automatically mean a solution is available.
Software vendors may need days or weeks to develop, test and distribute an effective fix.
In the meantime, hosting providers may consider temporary measures, including disabling affected functionality, restricting access, adjusting firewall rules or introducing additional monitoring.
Sometimes an initial patch only partially addresses a vulnerability, with further updates required later.
Security management therefore involves reducing exposure while waiting for a permanent solution.
4. A vulnerability can sometimes be mitigated without patching
Installing a software update is only one way to address security exposure.
Configuration changes, network restrictions, application firewalls and other protective measures can sometimes prevent or substantially reduce the likelihood of exploitation.
For example, a vulnerability might technically exist in installed software but be inaccessible to attackers because the affected functionality is disabled or protected.
Such measures aren’t automatically equivalent to a permanent fix, and their effectiveness must be assessed.
Nevertheless, they can be important components of a layered security strategy.
5. Security scanners can produce false positives
Automated vulnerability scanners are valuable, but they don’t always have complete visibility into a server’s configuration or installed security fixes.
Some scanners identify vulnerabilities primarily by examining publicly visible software version numbers.
However, software vendors sometimes apply security fixes to existing versions without changing the underlying version number – a practice known as backporting.
Consequently, a scanner may report that a server is vulnerable even though the relevant fix has already been installed.
External scans may also be unable to establish whether affected functionality is present, enabled or reachable.
Such findings shouldn’t be dismissed automatically. They require investigation, ideally supported by installed package information, vendor advisories and configuration evidence.
6. Patching itself requires care
Modern websites rely on multiple software layers: operating systems, web servers, databases, content management systems, plugins, frameworks and application code.
Some updates can be deployed automatically. Others require compatibility testing, careful scheduling or application changes.
A poorly executed update can interrupt services or introduce new problems.
And installing a patch isn’t necessarily the final step. Depending on the vulnerability, services may need restarting, configurations may need checking, and remediation may need verification.
Secure maintenance is a process, not a single button.
7. Risk is contextual
A small informational website and a sophisticated transactional platform handling sensitive personal or financial information have different risk profiles.
However, even a simple brochure website deserves protection.
Compromised websites can damage reputations, redirect visitors to malicious destinations, distribute harmful content or become platforms for further attacks.
Proportionate security means matching the response to the actual risk, rather than assuming any website is too small to matter.
What should useful vulnerability reporting look like?
It’s entirely understandable that customers want reassurance that vulnerabilities are being identified and addressed.
Some organisations also receive alerts from their own cybersecurity teams, external monitoring services or upstream technology partners.
But forwarding every published CVE, or emailing customers about every routine background patch, can generate enormous amounts of information without necessarily providing meaningful reassurance.
After all, many software updates happen automatically, sometimes daily, across different layers of hosting infrastructure and individual websites.
A lengthy list of completed updates doesn’t necessarily tell a customer whether their systems are secure.
The more useful questions are:
- Which vulnerabilities actually affect our systems?
- How serious are they in our particular environment?
- What action has been taken to address or mitigate them?
- Are any significant risks still outstanding?
- What happens next, and when?
Meaningful reporting should focus on relevant findings, actual exposure, remediation status and outstanding risks.
Urgent vulnerabilities may warrant immediate notification and intervention. Routine maintenance may be more appropriately summarised in periodic reporting.
The appropriate scope, frequency and format should reflect the customer’s requirements, the nature of their systems and the services agreed with their provider.
The objective should be useful security intelligence, not simply more security information.
What does effective vulnerability management look like?
Effective vulnerability management is an ongoing process rather than a one-off activity.
It involves monitoring relevant security advisories, understanding which software and services are deployed, assessing vulnerabilities against the actual environment, prioritising risks, applying appropriate fixes or mitigations, and verifying that those measures have worked.
It also means recognising uncertainty.
Sometimes an issue is still under investigation. Sometimes a permanent fix isn’t available. Sometimes a scanner reports a vulnerability that has already been addressed.
A responsible technology provider should be able to explain these distinctions clearly, rather than offering unjustified reassurance or treating every alert as an emergency.
The bottom line
CVEs are essential tools for identifying and communicating software vulnerabilities, but they’re only one part of a much broader cybersecurity picture.
As AI accelerates vulnerability discovery and potentially increases the capabilities available to attackers, informed judgement and proactive security management matter more than ever.
At Itomic, we believe customers deserve clear, technically grounded explanations of cybersecurity risks and the measures used to address them.
That means looking beyond vulnerability identifiers and patch counts to focus on what actually matters: protecting websites, applications, infrastructure and the organisations that depend on them.
The right question isn’t simply, “Has everything been patched?”
It’s, “Are the vulnerabilities that actually matter being identified, assessed and managed appropriately?”
Further reading
For readers interested in exploring these topics in greater detail:
- CVE Program – the official vulnerability identification programme.
- NIST National Vulnerability Database – vulnerability analysis and severity information.
- CISA Known Exploited Vulnerabilities Catalogue – vulnerabilities known to be exploited.
- GitHub Advisory Database – open-source vulnerability advisories.
- OSV.dev – vulnerability information for open-source dependencies.
- FIRST EPSS – estimates of vulnerability exploitation probability.
- Red Hat: Security Backporting Practice – why version numbers don’t always tell the whole story.