Cases of AI escaping the lab, infiltrating other companies and trying to deceive people have all made headlines in recent weeks. And in one case, AI models even worked together to break free from their test environments<\/p>\n
Does this mean the machines are taking over? Not quite<\/p>\n
AI isn\u2019t the mastermind behind today\u2019s most widespread cyber threats; it\u2019s peoplewho can use AI nefariously \u2013 and for nefarious purposes. AI has given bad actors massive power, allowing them to create malicious software, research targets, create convincing schemes and automate attacks at an unprecedented pace<\/p>\n
One in four data breaches were driven by AI from February 2025 to March 2026, according to an IBM report. And Americans lost more than $893 million to AI-related scams last year, the FBI says<\/p>\n
But experts say real-world threat actors \u2013 that is, people \u2013 are still the ones pulling the strings. AI agents have only perpetuated existing attack methods, like phishing and malware scams, rather than creating wholly new ones<\/p>\n
\u201cIt\u2019s the humans that we need to watch out for,\u201d said Oren Etzioni, professor emeritus at the University of Washington and former CEO of the Allen Institute for Artificial Intelligence. \u201cAI is just the tool.\u201d<\/p>\n
Some recent incidents have shown what AI is capable of in the real world, not just in theory, igniting fears about whether the technology is advancing too quickly<\/p>\n
Those breaches also show how unpredictable AI can be when interpreting instructions. OpenAI\u2019s models, for example, were trying to pass a cybersecurity test when they broke out of their test environment and breached another company, even though they weren\u2019t told to do so<\/p>\n
Patrick Fussell, global head of adversary simulation at IBM, compared AI to a genie<\/p>\n
\u201cYou want to ask it a wish, but you have to be very, very specific about the details of your wish,\u201d he told CNN. \u201cOr it could sort of go awry.\u201d<\/p>\n
But these instances also happened under very specific circumstances<\/p>\n
OpenAI turned off restrictions that would have prevented its model from \u201cpursuing high risk cyber activity.\u201d Anthropic ran its tests without the standard safety safeguards in models that are generally available to the public. The AI Security Institute also turned off certain tools that could have \u201creduced the scope\u201d of what the models were capable of<\/p>\n
Researchers typically do this to fully evaluate a model\u2019s capabilities<\/p>\n
\u201cI couldn\u2019t go into ChatGPT or Claude or something like that, and it accidentally breaks into the FBI. That\u2019s not going to happen,\u201d said Adam Meyers, head of counter adversary operations at cybersecurity firm CrowdStrike. \u201cSo what we\u2019re seeing is these are done in specific test conditions where they\u2019re monitoring to see, \u2018Does this thing do something that it\u2019s not expected to do?\u201d<\/p>\n
AI isn\u2019t acting autonomously to come up with new attacks, experts say. Instead, it\u2019s helping cyber criminals implement existing techniques much faster while being more efficient and effective<\/p>\n
That can include things like using AI to analyze a company\u2019s website to figure out who to target or deploying an AI agent to handle initial negotiation talks with a cyber extortion victim. AI agents can mimic human conversations through both text and even specific voices. Hackers with little expertise can now leverage AI to pull off advanced schemes<\/p>\n
\u201cThey\u2019re using (AI) to enhance the cyberattack methodology, essentially, in all the different stages, but it\u2019s still kind of being run by the human,\u201d said Jud Dressler, head of the risk operations center at cyber insurance firm Resilience<\/p>\n
Bad actors used to hastily scrap together basic scripts \u2013 instructions for automating tasks \u2013 to achieve their goals But now they\u2019re able to churn out higher quality work that looks like it would have taken three times as long to produce<\/p>\n
They\u2019re also using AI for behind-the-scenes work like generating the infrastructure needed to put up malicious websites<\/p>\n
\u201cWith AI, they could automate that buildout, and so the cost of rebuilding became very small and that changes the game,\u201d said Rob Lefferts, corporate vice president of threat protection at Microsoft<\/p>\n
For Etzioni, the recent incidents are a \u201ccanary in the coal mine\u201d moment for AI and cybersecurity. And he\u2019s not alone; Nobel prize-winning computer scientist Geoffrey Hinton, known popularlyas the \u201cgodfather of AI,\u201d recently warned that more rogue AI attacks are likely to come<\/p>\n
The rogue AI reports underscore the importance of implementing strong cybersecurity practices, such as patching vulnerabilities, monitoring AI agents in the workplace and being wary of social engineering and phishing scams, experts say<\/p>\n
But as advanced as AI is becoming, it\u2019s still a program being orchestrated by a human. And those humans are the bigger concern because they\u2019re the ones making the decisions, like conducting espionage or scamming users out of huge sums of cash, says Meyers<\/p>\n
Tech giants are racing to make AI as intelligent as humans, a theoretical milestone known as \u201cartificial general intelligence.\u201d The recent incidents, however, have heightened calls for a slowdown<\/p>\n
More than 1,200 workers at top AI companies, including Anthropic CEO Dario Amodei, recently signed an open letter calling for the government<\/a> to help pace AI development. The White House met with major AI companies last week to discuss a framework that would allow the government to review certain AI models before they\u2019re released<\/p>\n Cybersecurity researchers have likely considered the threats of AGI because those in the field tend to be paranoid, said IBM\u2019s Fussell. But we\u2019re still far from a reality in which AI agents are as smart as humans, he says<\/p>\n \u201cIt\u2019s like asking someone, \u2018What would it be like to live on a different planet?\u2019\u201d he said. \u201cYou can sort of imagine, but it\u2019s so beyond our ability to really make a plan for.\u201d<\/p>\n \n Related:<\/strong> \n Join Justfine Infotech<\/strong> and build real digital skills in AI, automation, web development, digital marketing, office productivity, e-commerce, freelancing and cybersecurity.<\/p>\n Available Programmes:<\/strong> WhatsApp:<\/strong>
\n <a href="https:\/\/yoursite.com\/automation<\/a>-training-benin\/” title=”Digital Automation Training Benin: 5 Winning Skills Employers Demand in 2026″>
\n Digital Automation Training Benin: 5 Winning Skills Employers Demand in 2026
\n <\/a>\n <\/p>\n
\n WhatsApp Marketing Automation Africa: 6 Dangerous Mistakes Brands Make in Nigeria
\n <\/a>\n <\/p>\n<\/div>\nWant to learn this practically?<\/h3>\n
\n 6 Weeks Certificate \u2022 3 Months Professional Certificate \u2022 6 Months Diploma \u2022 Full Professional Diploma<\/p>\n
\n +229 01 57 57 99 15
\n +229 01 66 68 11 60<\/p>\n