by Chatgpt can now deceive online review systems that are intended to prove that a user is human.
The latest version of the AI chatbot can serve as a personal assistant who is able to navigate on the web and perform tasks on behalf of the user.
Chatgpt Creator Openai describes this artificial intelligence of the next generation, which is known as Agentic AI, as a “natural development” of technology, with which AI can carry out actions such as online purchases, booking restaurants and the planning of appointments.
Early users of the new Chatgpt – there are currently only available in the Pro Plus and team versions – shared their experiences with the use of the Agentic AI and reveals that it is able to simply bypass security control points.
In an incident that was shared on Reddit, Chatgpt explained the process of dealing with the anti-bot check of Cloudflares.
“I click on the check box” I verify that you are “to complete the review of the cloud flare”, wrote the bot. “This step is necessary to prove that I am not a bot and can continue with the action.”
The Cloudflare system is one of the most common security measures used by locations to block automatic traffic. The check box “I am not a robot” is often used instead of a challenging captcha puzzle, although websites may have to re -evaluate your bot test methods.
Openaai said his latest chat bot would always apply for permission before they seize follow from episode and can be interrupted at any time.
“Chatgpt can now work for you with your own computer and do complex tasks from start to finish,” said the company in a blog post to present the new functions.
“Chatgpt navigently navigated on websites, filtered the results, asked them to safely register, carry out code, carry out analyzes and even provide fixed slide shows and spreadsheets that summarize its results.”
It is the first time that users can ask Chatgpt to take actions on the web, and similar starts of Chinese competitors as manus.
Openai recognized the risks that were granted a certain amount of autonomy with the AI. However, she claims to have strengthened its protective measures.
“We have strengthened the robust controls … and security measures for challenges such as the treatment of sensitive information in the live web, the broader user range and the (limited) terminal network access” added, “said the company.
“While these reductions significantly reduce the risk, the expanded tools of the chatt agent and the wider user reach of its overall risk profile means higher.”
