An intelligence alliance says AI-powered cyberattacks may be closer than anyone expected

Artificial intelligence has already changed the way businesses write code, analyze data, and automate tasks. Now, an international alliance of intelligence agencies says the technology could soon transform cyberattacks in ways that governments and companies may not be prepared for.

The Five Eyes intelligence alliance has issued a rare warning that advanced AI models capable of launching major cyberattacks could emerge within months rather than years. The group urged governments and businesses to strengthen their defenses immediately as cyber risks continue to evolve.

A stark warning from the world’s intelligence agencies

The Five Eyes alliance includes the United States, the United Kingdom, Canada, Australia, and New Zealand. In a joint statement, the agencies said frontier AI models are expected to dramatically change both offensive and defensive cyber capabilities.

The group warned that AI can lower barriers for malicious actors and increase the speed and complexity of attacks. According to the alliance, organizations should act now instead of waiting for these threats to become a reality.

The warning is notable because intelligence agencies do not often issue public statements about emerging technologies. Their message suggests that some of the world’s top security officials believe the threat is approaching much faster than many expected.

Why advanced AI models are causing concern

The warning comes amid growing attention on the capabilities of the latest AI systems. Researchers and technology executives have increasingly discussed the possibility that powerful models could help attackers identify vulnerabilities and automate complex cyber operations.

The Trump administration recently ordered AI company Anthropic to suspend access to its most advanced models for foreign nationals. The move followed concerns surrounding the company’s newest systems and their ability to discover security flaws.

Anthropic has said the government became aware of a method that could bypass safety protections in one of its public models. The company and government officials have reportedly been working together to address the issue.

Businesses may be the most vulnerable targets

Experts say sophisticated corporations are generally better prepared because they already invest heavily in cybersecurity. Smaller companies, however, may face greater risks because many have fewer resources and weaker defenses.

Olivia Shen, director of the Strategic Technologies Program at the United States Studies Center at the University of Sydney, said many organizations have significant gaps in their cyber protections. She warned that underprepared businesses could become easy targets if AI-powered attacks become more common.

The concern extends beyond governments and major corporations. Small and medium-sized businesses around the world may struggle to keep pace as cyber threats become faster and more sophisticated.

AI is both the problem and part of the solution

Despite the warning, intelligence officials also believe artificial intelligence can help strengthen cybersecurity. Organizations are increasingly using AI tools to detect vulnerabilities, monitor unusual activity, and respond to incidents more quickly.

The Five Eyes alliance said companies that integrate AI into their security operations could improve software quality and identify threats earlier. In other words, the same technology that could empower attackers may also provide defenders with new tools.

This dual nature of AI has complicated efforts to create regulations. Policymakers are trying to encourage innovation while also reducing the risks that come with increasingly capable systems.

Little-known fact: Cybercrime is projected to cost the global economy $10.5 trillion annually, making it one of the largest economic threats in the world.

The race to create guardrails continues

Independent assessments show that frontier AI models are rapidly improving at cyber tasks and, in controlled tests, some have completed multi-step simulated attacks. The rapid pace of development has made it difficult for lawmakers and regulators to establish clear rules.

The United States does not have a single comprehensive federal AI law, although voluntary frameworks, agency guidance, executive policy efforts, and state-level rules all shape how the technology is governed. Some experts argue that stronger government involvement is necessary, while others worry that excessive regulation could slow innovation.

Earlier this month, dozens of cybersecurity researchers, AI entrepreneurs, and executives signed an open letter calling for an open and scientific process for assessing AI risks. They argued that security teams need better tools to fix vulnerabilities before cybercriminals can exploit them.

Walking a difficult line

Experts say governments face a difficult balancing act. Artificial intelligence can dramatically improve cyber defenses, but it can also be used by criminals, scammers, and hostile actors.

Shen said there needs to be a set of ground rules that allows organizations to maximize AI’s defensive benefits while limiting its misuse. The challenge is creating those safeguards before the technology becomes even more powerful.

The Five Eyes warning suggests there may not be much time left to prepare. If advanced AI-driven cyberattacks are indeed only months away, governments and businesses may need to rethink their security strategies much sooner than expected.

TL;DR

• The Five Eyes intelligence alliance warned that powerful AI-driven cyberattacks could emerge within months.
• The alliance includes the United States, the United Kingdom, Canada, Australia, and New Zealand.
• Officials believe AI can make cyberattacks faster, more sophisticated, and easier to launch.
• Small and medium-sized businesses may be especially vulnerable because many lack strong cybersecurity defenses.
• Experts also see AI as part of the solution because it can help organizations detect and respond to threats more quickly.
• Policymakers are still struggling to create rules that balance innovation with security concerns.

This article was made with AI assistance and human editing.

Don’t forget to follow us for more exclusive content.

If you liked this, you might also like:

• Microsoft Chatbot can now design custom apps
• China targets long chatbot sessions with log-off nudges
• Perplexity Launches a Smarter Way to Research