Security teams embrace agentic AI
Companies and their cybersecurity teams are leaning into the new agentic world, experts say.
Why it matters: Agentic AI can reduce workload and boost response times, but if it misfires, it could expose systems to serious threats.
The big picture: While chatbots respond to prompts, agentic AI goes a step further and takes approved actions based on its own findings.
Like any technological evolution, getting security teams to adopt AI takes time and education.
Building confidence in new AI-enabled security tools also comes with a unique threat: If an AI tool gets something wrong, it leaves an opening for spies and cybercriminals to break in.
Driving the news: Microsoft unveiled plans Monday to start previewing 11 new AI agents in Security Copilot next month.
CrowdStrike added agentic AI to its security tools last month, and Trend Micro rolled out autonomous agents and its own AI brain to customers last year.
Flashback: Just two years ago, major corporations were blocking employees from even opening ChatGPT for fears of data leaks.
Yes, but: The tides have turned, and security is one of the clearest use cases for generative AI — especially since the industry has long had a dearth of available workers and faces high burnout rates.
65% of CISOs said in a survey last summer that their organizations are considered either "early adopters" or "early majority" adopters of new AI technologies, which could be influencing their newfound trust in AI tools.
Half of the CISOs in that same survey also said they have developed some AI use cases or were piloting potential new AI projects for their teams.
Between the lines: Many security teams just want agentic AI to help sort through the thousands of threat notifications they receive daily and determine which ones are legitimate threats to their organizations.
When Microsoft customers first started playing around with their Security Copilot, they would stick to prescriptive use cases, like summarizing a recent incident, Dorothy Li, corporate VP of Microsoft Security Copilot, told Axios.
As they've become more comfortable, some users now let Copilot automate as much of their workflow as possible, she added, which inspired Microsoft to bring autonomous agents into the mix.
Many of those use cases involved responding to phishing alerts and notifications about vulnerabilities across the various tools in their stacks.
Zoom in: Last month, CrowdStrike added an agentic capability to its security-focused large language model that automatically triages notifications for customers' security operations teams.
Once implemented, the new tool can eliminate more than 40 hours of manual work per week, CrowdStrike estimates.
CrowdStrike tests its new agentic capabilities internally against its own analysts' findings to ensure the tools are accurate and don't take inappropriate actions before they're deployed.
That testing is key to building trust with customers, who include security teams in major corporations, Elia Zaitsev, chief technology officer at CrowdStrike, told Axios.
"Everything in the generative AI space, in particular, by pretty much every measurement I've seen, is being adopted quicker than any technology out there," Zaitsev said.
Reality check: A healthy amount of skepticism still remains in AI's promise for security teams, Zaitsev added.
"People need to see those hard, quantifiable metrics," he said. "They need to see there's real ROI."
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
40 minutes ago
- Yahoo
Analysts unveil bold forecast for Alphabet stock despite ChatGPT threat
Analysts unveil bold forecast for Alphabet stock despite ChatGPT threat originally appeared on TheStreet. You typed in a question and clicked a few links, and Google could get paid if you landed on an ad. For years, that simple cycle helped turn Google into a trillion-dollar titan. But now, that model is under threat. 💵💰💰💵 AI-powered chatbots like OpenAI's ChatGPT are rapidly changing how people find answers. Instead of browsing through links, users are getting direct summaries on AI. These 'zero-click' searches quietly erode the economics that built the modern internet. The number of users is growing fast. OpenAI CEO Sam Altman said in April that ChatGPT already has 'something like 10% of the world" in terms of users, pegging the number closer to 800 million, Forbes reported. Even Google seems to know it. It's giving AI answers, called AI Overviews, right at the top of the page. "What's changing is not that fewer people are searching the that more and more the answers to Google are being answered right on Google's page. That AI box at the top of Google is now absorbing that content that would have gone to the original content creators," Cloudflare CEO Matthew Prince said in a CNBC interview. Alphabet () , Google's parent company, isn't showing any cracks just yet. In April, the company posted first-quarter revenue of $90.23 billion, topping Wall Street expectations. Earnings per share came in at $2.81, far above the forecasted $ the backbone of Google's business, brought in $66.89 billion, accounting for nearly three-quarters of total revenue. Its 'Search and other' segment rose almost 10% year over year, hitting $50.7 billion. Meanwhile, Google's own AI tools are starting to show traction. AI Overviews now has 1.5 billion users per month, up from 1 billion in October, the company said. So far, the numbers suggest that AI isn't cannibalizing Google's business yet. Bank of America remains bullish on Alphabet stock. The firm reiterated a buy rating and a price target of $200, which implies a potential 15% upside from current levels, according to a recent research report. The firm said in May, Google's global average daily web visits held steady at 2.7 billion, unchanged from the previous month and down 2% from a year earlier. ChatGPT, meanwhile, saw a 3% month-over-month increase to 182 million, marking a 105% jump the U.S., Google traffic slipped 2% year-over-year to 524 million daily visits, while ChatGPT surged 112% over the same period to 26 million. Although Google has highlighted the growing reach of its AI Overviews, analysts are uncertain whether it's translating into more traffic. 'So far, we are not seeing a lift in Google traffic from AI Overviews expansion, though we think the search experience is much improved,' the analysts wrote. The competition is real. Google's global search share also edged down in May, falling 8 basis points month-over-month and 123 basis points year-over-year to 89.6%, according to Statcounter. Still, Bank of America analysts remain optimistic on Alphabet stock. "While ChatGPT's traffic continues to grow rapidly, we think Google remains well-positioned given its scale, multi-product reach, data assets, and robust monetization infrastructure," the analysts said. "AI can expand overall search monetization by better understanding the intent behind complex and long-tail queries that were previously hard to monetize," they added. Morningstar's Malik Ahmed Khan echoed that sentiment, saying Alphabet's diverse revenue streams and global exposure should cushion any hits, even as regulatory and AI risks mount, according to a May research report. Alphabet stock closed at $174.92 on June 6. The stock is down 8% unveil bold forecast for Alphabet stock despite ChatGPT threat first appeared on TheStreet on Jun 6, 2025 This story was originally reported by TheStreet on Jun 6, 2025, where it first appeared. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
The Verge
2 hours ago
- The Verge
'You can now easily pause, play, skip forward or backward, and seek via the progress bar right from the picture-in-picture window,'
Microsoft Edge's picture-in-picture mode is getting better controls. according to a blog post. I don't use Edge, but these seem like good upgrades.
Yahoo
2 hours ago
- Yahoo
Inside the Secret Meeting Where Mathematicians Struggled to Outsmart AI
On a weekend in mid-May, a clandestine mathematical conclave convened. Thirty of the world's most renowned mathematicians traveled to Berkeley, Calif., with some coming from as far away as the U.K. The group's members faced off in a showdown with a 'reasoning' chatbot that was tasked with solving problems they had devised to test its mathematical mettle. After throwing professor-level questions at the bot for two days, the researchers were stunned to discover it was capable of answering some of the world's hardest solvable problems. 'I have colleagues who literally said these models are approaching mathematical genius,' says Ken Ono, a mathematician at the University of Virginia and a leader and judge at the meeting. The chatbot in question is powered by o4-mini, a so-called reasoning large language model (LLM). It was trained by OpenAI to be capable of making highly intricate deductions. Google's equivalent, Gemini 2.5 Flash, has similar abilities. Like the LLMs that powered earlier versions of ChatGPT, o4-mini learns to predict the next word in a sequence. Compared with those earlier LLMs, however, o4-mini and its equivalents are lighter-weight, more nimble models that train on specialized datasets with stronger reinforcement from humans. The approach leads to a chatbot capable of diving much deeper into complex problems in math than traditional LLMs. To track the progress of o4-mini, OpenAI previously tasked Epoch AI, a nonprofit that benchmarks LLMs, to come up with 300 math questions whose solutions had not yet been published. Even traditional LLMs can correctly answer many complicated math questions. Yet when Epoch AI asked several such models these questions, which were dissimilar to those they had been trained on, the most successful were able to solve less than 2 percent, showing these LLMs lacked the ability to reason. But o4-mini would prove to be very different. [Sign up for Today in Science, a free daily newsletter] Epoch AI hired Elliot Glazer, who had recently finished his math Ph.D., to join the new collaboration for the benchmark, dubbed FrontierMath, in September 2024. The project collected novel questions over varying tiers of difficulty, with the first three tiers covering undergraduate-, graduate- and research-level challenges. By February 2025, Glazer found that o4-mini could solve around 20 percent of the questions. He then moved on to a fourth tier: 100 questions that would be challenging even for an academic mathematician. Only a small group of people in the world would be capable of developing such questions, let alone answering them. The mathematicians who participated had to sign a nondisclosure agreement requiring them to communicate solely via the messaging app Signal. Other forms of contact, such as traditional e-mail, could potentially be scanned by an LLM and inadvertently train it, thereby contaminating the dataset. The group made slow, steady progress in finding questions. But Glazer wanted to speed things up, so Epoch AI hosted the in-person meeting on Saturday, May 17, and Sunday, May 18. There, the participants would finalize the final batch of challenge questions. Ono split the 30 attendees into groups of six. For two days, the academics competed against themselves to devise problems that they could solve but would trip up the AI reasoning bot. Each problem the o4-mini couldn't solve would garner the mathematician who came up with it a $7,500 reward. By the end of that Saturday night, Ono was frustrated with the bot, whose unexpected mathematical prowess was foiling the group's progress. 'I came up with a problem which experts in my field would recognize as an open question in number theory—a good Ph.D.-level problem,' he says. He asked o4-mini to solve the question. Over the next 10 minutes, Ono watched in stunned silence as the bot unfurled a solution in real time, showing its reasoning process along the way. The bot spent the first two minutes finding and mastering the related literature in the field. Then it wrote on the screen that it wanted to try solving a simpler 'toy' version of the question first in order to learn. A few minutes later, it wrote that it was finally prepared to solve the more difficult problem. Five minutes after that, o4-mini presented a correct but sassy solution. 'It was starting to get really cheeky,' says Ono, who is also a freelance mathematical consultant for Epoch AI. 'And at the end, it says, 'No citation necessary because the mystery number was computed by me!'' Defeated, Ono jumped onto Signal early that Sunday morning and alerted the rest of the participants. 'I was not prepared to be contending with an LLM like this,' he says, 'I've never seen that kind of reasoning before in models. That's what a scientist does. That's frightening.' Although the group did eventually succeed in finding 10 questions that stymied the bot, the researchers were astonished by how far AI had progressed in the span of one year. Ono likened it to working with a 'strong collaborator.' Yang Hui He, a mathematician at the London Institute for Mathematical Sciences and an early pioneer of using AI in math, says, 'This is what a very, very good graduate student would be doing—in fact, more.' The bot was also much faster than a professional mathematician, taking mere minutes to do what it would take such a human expert weeks or months to complete. While sparring with o4-mini was thrilling, its progress was also alarming. Ono and He express concern that the o4-mini's results might be trusted too much. 'There's proof by induction, proof by contradiction, and then proof by intimidation,' He says. 'If you say something with enough authority, people just get scared. I think o4-mini has mastered proof by intimidation; it says everything with so much confidence.' By the end of the meeting, the group started to consider what the future might look like for mathematicians. Discussions turned to the inevitable 'tier five'—questions that even the best mathematicians couldn't solve. If AI reaches that level, the role of mathematicians would undergo a sharp change. For instance, mathematicians may shift to simply posing questions and interacting with reasoning-bots to help them discover new mathematical truths, much the same as a professor does with graduate students. As such, Ono predicts that nurturing creativity in higher education will be a key in keeping mathematics going for future generations. 'I've been telling my colleagues that it's a grave mistake to say that generalized artificial intelligence will never come, [that] it's just a computer,' Ono says. 'I don't want to add to the hysteria, but in many ways these large language models are already outperforming most of our best graduate students in the world.'
