AI Is Becoming a Hacker’s Tool: Businesses Need to Rethink Cybersecurity

Artificial intelligence is no longer just a defensive tool. It is increasingly being used on the other side — to accelerate real-world cyberattacks.
In August 2026, a striking case came to light: the Russian-speaking hacking group Aur0ra used the Cursor AI agent in attacks against at least seven companies.
Researchers discovered logs of the attackers’ interactions with the AI and found that malicious tasks were presented to the agent as legitimate security testing.
According to Reuters, AI assisted the attackers in identifying vulnerabilities, working with credentials, and carrying out other stages of the attacks.
When the system refused suspicious requests, the attackers reportedly attempted to bypass its safeguards by starting new sessions and again presenting their activities as authorized security testing.
This is not an isolated case. More than 100 technology and financial companies, including OpenAI, Microsoft, Alphabet, Amazon, and IBM, have called for stronger cyber defenses as the threat of AI-enabled attacks continues to grow.
Why does this matter for businesses?
The biggest change is not necessarily the emergence of entirely new vulnerabilities.
AI makes existing attacks significantly faster.
According to Verizon’s 2026 Data Breach Investigations Report, vulnerability exploitation became the leading initial access vector for the first time in the report’s 19-year history, accounting for 31% of breaches.
AI can also dramatically reduce the time between discovering a known vulnerability and exploiting it — potentially shrinking the defensive window from months to just hours.
What can businesses do now?
• Apply the principle of least privilege. Employees, applications, and AI agents should only have access to the data and systems they genuinely need.
• Require MFA. This is particularly important for administrator accounts, corporate email, VPNs, cloud services, and critical infrastructure.
• Patch vulnerabilities faster. If attackers can exploit weaknesses more quickly, traditional patching timelines may no longer be sufficient.
• Control corporate AI usage. Unknown GPT services, plugins, and AI agents should not have unrestricted access to corporate documents, source code, passwords, tokens, or API keys.
• Log AI-agent activity. Critical operations — changing permissions, executing code, deleting data, or performing actions in production environments — should require additional controls or human approval.
• Conduct regular security assessments. Organizations should test not only traditional attack scenarios but also attacks enhanced or automated by AI.
What about the legal side?
AI security is no longer solely an IT department responsibility.
Companies should establish internal policies governing the use of AI, define permitted access levels for AI agents, clarify employee responsibilities, set requirements for third-party AI providers, and develop incident-response procedures specifically addressing AI-related security events.
Because the next cyberattack may not simply happen faster.
Part of the attacker’s work may be performed by a machine.
So the question for businesses is no longer:
“Can hackers use AI?”
They can.
The real question is whether your defenses can respond faster.
SynCo — Cyber Legal Bureau Protecting businesses where technology meets risk.
Official source
SYNCO
Open on lex.uzThis material was prepared by SYNCO lawyers based on the legislation of the Republic of Uzbekistan and is for information only. It is not legal advice — please consult a specialist regarding your situation.

