An e-commerce earthquake is coming. It will completely transform the industry.

• OpenAI has launched a new service called ChatGPT Agent.
• The company's technology CEO, Sam Altman, announced the new service and urged caution when using it.
• OpenAI's new offering is causing a stir among artificial intelligence experts and e-commerce industry representatives. According to InPost CEO Rafał Brzoska, services like ChatGPT Agent will "redefine the entire market."

As we read on the OpenAI website, the new service allows ChatGPT to handle requests such as "look at my calendar and inform me about upcoming client meetings based on the latest news," "plan and buy ingredients for a Japanese breakfast for four people," and "analyze the activities of three competitors and create a presentation."

ChatGPT will navigate websites, filtering results, requesting logins and accesses as needed , run code, perform analysis, and even provide editable slideshows and spreadsheets summarizing the results.

At the heart of this new functionality is a unified agent system. It combines the three strengths of previous breakthrough solutions: the operator's ability to interact with websites, the deep research capability to synthesize information, and the intelligence and conversational fluency of ChatGPT.

Sam Altman: We don't know exactly what the consequences will be

ChatGPT performs these tasks using its own virtual computer, switching between reasoning and action to handle complex workflows from start to finish, all based on user instructions.

"The agent represents a new level of AI capabilities and can perform extraordinary, complex tasks for you, using its own computer. It combines the spirit of Deep Research and Operator (the company's earlier products - ed.), but it's more powerful than you might think - it can think long-term, use certain tools, think more, take action, think more, and so on. For example, during the launch, we showed a demo of preparing for a friend's wedding: buying an outfit, booking travel, choosing a gift, and so on. We also showed an example of data analysis and creating presentations for work," Sam Altman, CEO of OpenAI, wrote on the X platform.

"We don't know exactly what the consequences will be, but malicious actors could try to trick users' AI agents into giving them private information they shouldn't and taking actions they shouldn't, in ways we can't predict. We recommend granting agents the minimum access required to complete their tasks to reduce the risk of privacy and security breaches," Altman wrote.

Rafał Brzoska: the e-commerce revolution is happening before our eyes

OpenAI's new service is already generating a lot of buzz among artificial intelligence experts and e-commerce industry representatives. According to InPost CEO Rafał Brzoska, services like ChatGPT Agent will "redefine the entire market."

"For those who haven't seen it yet, the e-commerce model of buying on traditional marketplace sites will end sooner than you think. Agentic AI shopping will redefine the entire market—whether anyone likes it or not. Content publishers have already experienced this, with traffic from search engines like Google dropping by around 60% in some cases in the last three months. A similar wave will come to online shopping. The next 12-18 months in our industry are shaping up to be interesting," Brzoska wrote on the X platform on Wednesday.

A day later, after the premiere of the new service from Open AI, Brzoska emphasized that it would be "a revolution happening before our eyes."

"(...) Overnight, OpenAI launches a new ChatGPT with a highly advanced Operator mode, which effectively becomes a BUYER tool that fulfills purchase orders FOR US AND FOR US. Time savings for those who, instead of poring over online stores and searching for products, define a precise task that the bot will "deliver" with the purchase. The e-commerce revolution is happening before our eyes and will redefine markets," Rafał Brzoska wrote on the X platform on Thursday.

The more autonomous the agent, the more access rights it will need.

"If a user wants an AI agent to find and buy a dress for them without asking any further questions, in addition to internet access, the AI agent will need access to their wallet. If a user wants the agent to plan an event and invite friends, they will need access to at least their calendar and contact list," Luiza Javorsky, co-founder of AI, Tech & Privacy Academy, wrote on LinkedIn. "Privacy threats associated with the ChatGPT agent will grow exponentially as many people grant access to external tools containing personal data (calendar, email, wallet, and more). OpenAI recognizes that malicious agents will attempt to trick other people's AI agents into sharing their private information," she added.

Unfortunately, the pace of AI development is much faster than the pace of development of knowledge about AI, wrote Luiza Javorsky.

The ChatGPT agent is equipped with a number of tools

OpenAI has equipped the ChatGPT agent with a set of tools:

• a visual browser that communicates with the network via a graphical user interface,
• a text browser for simpler reasoning-based web queries, a terminal and direct access to the API (application programming interface, i.e. rules and protocols enabling communication between applications and systems).
• The agent can also leverage ChatGPT connectors, which allow it to connect apps like Gmail and GitHub so that ChatGPT can find information relevant to the user's queries and use it in its responses.

"We believe it's important to start learning from experience and for people to carefully and slowly implement these tools as we become better at quantifying and minimizing potential risks. As with other new levels of opportunity, society, technology, and risk mitigation strategies will need to co-evolve," Sam Altman wrote on X.