A new survey paper from Tencent and several Chinese universities argues that today's AI is still stuck in the role of a clever assistant rather than a genuine "digital colleague," according to The Decoder.
The core claim is straightforward: AI systems won't become reliable coworkers as long as they simply generate answers. To earn that label, the researchers say, they need to finish entire tasks from start to end.
That shift hinges on where and how the AI operates. According to the paper, the key is having AI work inside persistent work environments — settings that stick around across a job — instead of producing a one-off reply and stopping there.
The distinction matters because most people experience AI today as a chatbot: you ask, it responds, and the follow-through is left to you. A tool that drafts an email is useful, but a coworker is someone who actually completes the assignment. The researchers frame this as the path from chatbot to colleague.
The paper is a survey, meaning it maps out and synthesizes the broader trajectory of the field rather than announcing a single new product or breakthrough.
Why it matters: how researchers define an "AI coworker" will shape what the next generation of AI tools is built to do — and whether they end up doing your work or just talking about it.