AI can help organizations detect threats faster. But it also enables cybercriminals to launch more sophisticated, scalable, and evasive attacks.

Below, we break down the latest data around AI-driven cyber attacks and what it all means for organizations moving forward.

Contents

Top AI Cyber Attack Stats

What Percentage of Cyberattacks Use AI?

A 2025 report revealed that 82.6% of phishing emails are created using AI.

That’s a significant increase of 53.5% over the previous year.

Source: KnowBe4

AI-Driven Threat Statistics

Nearly 9 in 10 organizations (87%) have been the targets of AI-powered cyberattacks over the last 12 months.

And that problem is predicted to worsen, with 91% of security experts anticipating an increase in AI-driven threats through 2028.

AI-powered cyberattacks can increase costs, but they also present a business opportunity for cybersecurity pros. As of 2025, the global AI cybersecurity market is worth an estimated $29.64 billion.

That figure is expected to rise to $167.77 billion by 2035, growing at a CAGR of 18.93%.

With the scale and frequency of cyber threats growing, 69% of enterprises claim that AI has become a necessary component within their cyber defenses.

AI also has the power to help prevent and mitigate security threats. According to a recent survey, organizations can boost their cyber threat detection by 1.6x by integrating AI into their cybersecurity stack.

Sources: SoSafe, Deloitte, Precedence Research, Industrial Cyber

AI Phishing Stats

With the emergence of AI in cyber attacks, it’s no surprise to see a growing number of phishing scams.

In the second half of 2024, there was a reported 202% surge in phishing emails.

The proliferation of ready-made phishing kits created by AI is leading to a significant rise in credential phishing attacks.

In the second half of 2024, there was a rise of 703% in these sorts of attacks.

And according to a recent study, 3 in 5 participants were fooled by AI-automated phishing. Experts state that this success rate is comparable to human-generated phishing scams. However, AI allows for a massively increased scale.

Over half of all AI-based phishing scams receive a click-through.

By comparison, phishing content produced manually results in a 12% click-through rate. AI phishing’s click-through rate is over 4x better than human content.

Concerningly, phishing scams can now be launched for a fraction of the price. The phishing process can be wholly automated with LLMs, cutting costs by approximately 95% when compared to traditional methods.

What’s more, Hackers are now able to compose phishing scams at a much faster rate using generative AI. In some cases, this can be up to 40% quicker.

Sources: SlashNext, Harvard Business Review, CrowdStrike, SoSafe

Deepfake Cyber Attack Statistics

According to McAfee, the average American comes across 2.6 deepfakes daily. This figure increases to 3.5 per day for young adults aged 18 to 24.

This is especially worrying as the same report reveals that 4 in 5 people struggle to identify a deepfake from genuine content, paving the way for deepfake scams.

Deepfakes are commonly found on social media platforms, with Meta (49%) and TikTok (47%) among the worst offenders.

As a result, concern over AI and its role in deepfake scams is growing. The latest findings show that 63% of cybersecurity leaders are uneasy about it.

And with good reason. The number of deepfake incidents in Q1 2025 surpassed all of 2024.

But deepfake incidents, which include fraud, have been on the rise for years.

Between 2017 and 2022, just 22 incidents were recorded. In 2023, this figure almost doubled, reaching 42. And in 2024, there were 150.

However, in Q1 2025 alone, 179 separate deepfake incidents were recorded – 19% more than the entirety of 2024.

At the same time, the number of deepfakes now makes up 6.5% of all fraud attacks. That’s a 2,137% rise between 2022 and 2025.

The financial sector is often targeted by deepfake attacks. Approximately 53% people in the finance industry have experienced a deepfake attack.

That’s likely because deepfake awareness has a long way to go. Around 1 in 5 (22%) of people do not know what a deepfake is. While 99.9% of people cannot reliably identify deepfakes.

To date, the biggest single financial loss from a deepfake scam totaled $25 million. In the well-publicized 2024 incident, a Hong Kong financial organization was deceived when AI was used to mimic the firm’s CFO.

Sources: McAfee, iProov, Egress, Surfshark, Signicat, Medius, iProov, CNN

AI Voice Cloning & Vishing Statistics

As of 2026, AI voice cloning is valued at an estimated $4.4 billion worldwide.

By 2033, the market is expected to expand by over 7x, reaching $25.6 billion.

AI voice scams are becoming more commonplace, with 10% of people being targeted and 15% knowing of someone who has been affected.

Although those figures may seem low, 77% of AI voice scam victims have consequently lost money.

One reason for the growing number of victims might be that most adults share their voice data online each week.

On average, 53% of adults share their voice online at least once a week, whether that’s through social media posts, comments, reviews, or other forms of digital expression. And 49% share this information as many as 10 times per week.

This highlights not only the scale of participation in online discourse but also the frequency with which people contribute their voice to online tasks and/or create content.

Alarmingly, people are only able to correctly recognize an AI-generated voice 60% of the time. One study found that, in 80% of cases, people believed an AI voice to be the same as the real voice.

Over 60s are the most common victims of voice cloning scams. Recent data from the Federal Trade Commission reveals that people over 60 years old are 45% more likely to fall victim to AI voice scams than those below that age.

Sources: market.us, McAfee, Nature, FOX 59

Future Outlook for AI Cyber Attacks

Cybercrime is predicted to increase by 50% between 2024 and 2028. As AI continues to pose new threats, the overall cost of cybercrime is set to grow to $13.82 billion by 2028.

Very few business/cyber leaders see AI as a net positive when it comes to cybersecurity. In fact, just 9% think that AI can give cybersecurity an advantage over attacks.

Around 1 in 3 (35%) don’t expect an advantage either way, while 56% think cybercriminals will have an advantage thanks to generative AI.

To prepare for the evolving landscape of AI-driven threats, pursuing cybersecurity degree programs can equip professionals with the knowledge and practical skills needed to defend against advanced attacks and stay ahead in the field.

Around 70 countries are set to implement AI-related policies to help address cybersecurity concerns.

Over 50 countries have already taken concrete steps to address growing public concerns about AI safety. These efforts reflect a global recognition that while AI holds immense potential, it also introduces risks that must be managed carefully.

To date, thousands of AI-related policies and initiatives have been proposed worldwide, spanning everything from national strategies to highly targeted sectoral regulations. These measures often focus on improving cybersecurity, ensuring ethical use, and reducing systemic risks.

Governments have created specialized AI councils, inter-agency task forces, and advisory boards, while legislatures are drafting legal frameworks designed to establish accountability, promote transparency, and safeguard against misuse.

Sources: SoSafe, AlixPartners, Mind Foundry

Conclusion

AI is transforming cyberattacks. As generative models become more accessible, even low-skill attackers can deploy advanced tools that once required specialized knowledge.

Organizations that fail to adapt their security strategies to account for AI-assisted threats will find themselves increasingly vulnerable.