{
  "id": 15604228,
  "title": "800,000 attacks a day: Is your AI protecting you or making you more vulnerable?",
  "url": "https://urgent.news/2026/10/19/800-000-attacks-a-day-is-your-ai-protecting-you-or-making-you-more",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-19T12:31:12.000Z",
  "source": {
    "name": "Gulf News",
    "slug": "gulf-news",
    "url": "https://gulfnews.com/technology/800000-attacks-a-day-is-your-ai-protecting-you-or-making-you-more-vulnerable-1.500716089"
  },
  "original_language": "en",
  "account": "October marks Cybersecurity Awareness Month in the UAE, where cyber threats have surged with more than 800,000 attempted attacks each day, as reported by the UAE Cyber Security Council. Security experts warn that the rise of autonomous AI agents, which make decisions, access data, and operate across company systems, is increasing vulnerabilities without sufficient human oversight. Early this year, UAE infrastructure faced between 90,000 and 200,000 daily breach attempts, which have since more than doubled to over 800,000. The Council documented 128 confirmed incidents, including ransomware, data leaks, and website defacement. According to Positive Technologies, the UAE accounts for 35% of all attacks in the region.\n\nDr. Mohamed Al Kuwaiti, the Council's head, attributes this surge to AI's acceleration of cyberattacks. Global cyber losses in 2024 reached an estimated $10.5 trillion, with projections suggesting a staggering $15.6 trillion by 2029. Industry leaders emphasize that agentic AI, which operates independently, introduces new entry points and targets for cybercriminals. Salah Suleiman, Managing Director at TrendAI, notes that AI agents can act as distinct identities, entry points, and targets.\n\nTrendAI's four-principle approach to securing agentic AI includes visibility, intent, control, and oversight. A dedicated AI security layer is proposed to govern model and agent behavior. Alexandre Depret-Bixio, SVP at Anomali, frames the issue as a boardroom concern, asking what organizations permit machines to decide on their behalf. Data quality is critical, as flawed data can lead to \"confident but bad\" AI decisions, posing unacceptable risks.\n\nGovernance remains a significant challenge. Only 47% of organizations, according to DXC Technology research, have fully integrated governance for agentic AI. Many are already deploying autonomous tools without adequate oversight, rapidly increasing risk. Daniel Churches, Director of Cybersecurity at DXC, urges companies to shift from reactive defense to proactive, autonomous security. Alexandre Depret-Bixio argues that autonomy can be trusted incrementally, with human approval and auditing for critical decisions.\n\nJustin Henkel, CISO at SolarWinds, describes machine identities as an \"enterprise control problem,\" not merely an identity access management issue. He urges leaders to enforce strict protocols, such as privilege reduction, owner coverage, and credential lifetime management to ensure trustworthy AI operations. The consensus is clear: attackers currently leverage AI at scale, and the UAE bears this burden daily. Companies must deploy AI agents more cautiously than they govern them, as failing to do so may invite criminal infiltration. As Depret-Bixio aptly puts it, \"The near-term future of AI will be judged by whether we can stand behind every decision it makes.\"",
  "summary": "Dubai: October is remembered as cybersecurity awareness month and now more than ever, it has become imperative for people to know more. Especially in the UAE where we’ve seen more than 800,000 attempted cyberattacks a day, according to the UAE Cyber Security Council. Security executives are now flagging newer risks. They point to the excessive use of autonomous AI agents who are taking decisions,…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}