Artificial Intelligence powerhouse, ChatGPT, has expanded its capabilities following an update, allowing it to perform tasks and make decisions on behalf of users.
**New ChatGPT Agent Mode Offers Comprehensive Assistance with Caution**
OpenAI has unveiled a new feature for its popular chatbot, ChatGPT, known as the "ChatGPT agent." This feature allows the chatbot to execute actions on a user's behalf, such as helping to prepare for a wedding or planning and buying ingredients for a meal. However, it's important to note that this new functionality comes with a word of caution from OpenAI CEO Sam Altman, who advises users to exercise caution when giving ChatGPT access to personal information.
The agent mode builds on and combines capabilities from the ChatGPT Operator and Deep Research tools already offered by OpenAI. Users can issue commands such as "look at my calendar and brief me on upcoming client meetings based on recent news" or "plan and buy ingredients to make Japanese breakfast for four." Certain tasks, such as sending an email, require user oversight in ChatGPT's agent mode.
ChatGPT operates across OpenAI's ecosystem, but lacks the native, deeply integrated cross-platform agent capabilities seen with Google's Gemini in Google Workspace, Pixel devices, and mobile apps. While ChatGPT is highly fluid in handling voice and images, with video input support reportedly on the roadmap but not fully live yet, Gemini 2.5 Pro supports text, image, audio, and video inputs, making it the most versatile multimodal model available.
Google's similar ambitions with its Gemini helper have also been announced, with the AI-powered agent offering advanced capabilities, notably in Google Search, where users can leverage AI to call local businesses automatically, gather pricing and availability, and receive information via text. This goes beyond passive Q&A, enabling the AI to act as a true digital agent.
However, the integration of Gemini into Google's ecosystem means it can access and act on a wide range of personal data—emails, Drive files, Photos, search history—for personalized, contextual responses. This integration is a core feature, enabling powerful agent-like functions, but it also raises significant privacy concerns, especially regarding consent, data retention, and potential misuse.
The new feature from OpenAI is available for those who subscribe to a Pro, Plus, or Team plan. Meanwhile, Google made AI-related announcements during its developer conference in May, including an agent for various tasks. Apple is also working on a more advanced version of Siri that can use apps on a user's behalf, although the update is delayed indefinitely.
It's important to remember that AI models, including ChatGPT, are still prone to hallucinations and bias, as demonstrated by xAI's Grok chatbot. As the industry continues to push for changes in the way people get things done on the Internet, it's crucial to approach these new technologies with a balanced understanding of their potential benefits and risks.
The ChatGPT agent, now available for subscribers of OpenAI's Pro, Plus, or Team plan, incorporates artificial-intelligence technology to help manage business tasks, such as preparing for a meeting by providing a brief based on recent news or planning and purchasing ingredients for a meal. This tech-driven assistance, however, should be used with caution, considering the risk of potential biases or hallucinations in AI models as demonstrated by xAI's Grok chatbot.
The advancements in artificial-intelligence technology, visible in the features of both ChatGPT and Google's upcoming AI agent, have the potential to revolutionize the tech industry, offering capabilities beyond passive question-answering and acting as true digital assistants. However, the deployment of AI-powered agents across various platforms raises concerns about privacy and consent, as these agents may have access to personal data and could potentially be misused.
