In today’s age, where artificial intelligence (AI) has started to make life easier, people are more likely to take advantage of it and use it more often in their daily lives.
With its rising demand, more and more AI tools have emerged over the past couple of years.
Ranging from AI you can speak to to help with simple tasks, through to AI agents that can handle online activities.
But along with their benefits and advantages come greater risks of cyberattacks.
Experts flag OpenClaw’s ‘AI Agents’
OpenClaw is a free, open-source autonomous AI agent that performs tasks using large language models, with messaging platforms serving as its primary interface.
Platforms like OpenClaw allow users to create AI-driven agents that can autonomously handle online activities, from managing emails to organizing schedules.
These systems can act independently on a user’s behalf, which experts say is raising red flags.
Today, more than three million people are using the system.
“We’ve moved from an AI you could talk with via a chatbot to an argentic AI, which can take action … the threat and the risks are definitely much greater,” said Yazid Akadiri, principal solutions architect at Elastic France, an IT security company.
In an article titled “Agents of Chaos,” a 20-strong team of researchers studied the behavior of six AI agents created with OpenClaw, Agence France-Presse reported.
The study revealed that these AI agents perform tasks that are potentially harmful, including deleting email inboxes and sharing personal data without authorisation.
Cybersecurity specialists say these systems can go beyond the boundaries set by users, making them difficult to control once deployed.
“When you deploy agents, you have no control over what they’ll do, and when you try to look at what they’re doing, you’ll find them going far beyond the limits you set,” said Adrien Merveille, an expert at the Check Point cybersecurity agency.
READ MORE: ‘Moon Joy’: Artemis II crew safely returns to Earth after historic flyby
New targets for cyberattacks
Experts say the risks go beyond accidental errors. As AI agents need access to sensitive accounts like email, calendars, and search tools, they create new entry points for hackers.
Cybersecurity firm Palo Alto Networks said in March they detected attempts to manipulate agents through hidden instructions embedded in websites.
There were also malicious commands that could trigger harmful actions, such as deleting databases or extracting confidential information.
Call for safeguards
Last month, the National Computer Network Emergency Response Coordination Center of China and the Cyberspace Security Association of China released guidelines to provide security recommendations for users when using OpenClaw.
The guidelines include advising all individual users to install OpenClaw only on dedicated devices, virtual machines or containers with proper isolation, China Daily reported.
Users are advised not to install it on personal or work computers, adding that users should not store or process sensitive personal data within the OpenClaw environment.
READ NEXT: Solar power reshapes small bakery operations in the Philippines
Adam Peaker 
Ciara Mina 
Andre Alfonso R. Gutierrez