On Thursday, OpenAI unveiled its new AI agent tool named Operator, introducing it as a research preview. However, the rollout has encountered some challenges.
The company demonstrated the capabilities of Operator during an online presentation, describing it as a Computer Using Agent (CUA) that operates on the GPT-4o model. This tool is designed to perform multi-modal tasks, including web searches and comprehending the logic behind the search results.
Despite its innovative features, early testers have raised concerns, highlighting issues such as slower performance than what was showcased in the demonstration and instances of misinformation similar to those found in the standard ChatGPT chatbot, as reported by Quartz.
Additionally, Benzinga noted that user complaints caught the attention of OpenAI’s CEO, Sam Altman. One user reported issues with the tool on a news website, prompting Altman to pledge a swift resolution. The user’s experience may have been influenced by hallucinations inherent in AI operation.
While many are intrigued by the functionalities showcased in OpenAI’s demo, the price of $200 per month for access under the ChatGPT Pro tier poses a significant barrier for many potential users.
Chris Smith from BGR mentioned that although he is a subscriber to ChatGPT Plus, he finds it difficult to justify the monthly cost for the Operator tool. However, OpenAI plans to eventually roll out Operator to its ChatGPT Plus, Team, and Enterprise plans.
One of the prevalent criticisms from users is that Operator is currently limited to the U.S., leaving European users feeling frustrated due to their inability to access the tool.
According to ComputerWorld, AI agents like Operator could present unique security risks, such as facilitating automated attacks or circumventing CAPTCHA protections. Although OpenAI claims to have robust security measures in place for Operator, experts have raised concerns about how this technology might conflict with existing data handling systems used by search engines like Google.
