Fulcrum Echo promises AI text in the style of any writer, so I tested it
Fulcrum, an AI startup founded in 2025, has released Echo, a language model designed to generate text in the style of any writer. The technology has prompted questions about the ethics of AI systems that can convincingly imitate individuals' distinctive voices, from celebrated authors to ordinary journalists.
A journalist from The Register tested Echo by providing one of their own articles and requesting a response to a question written in their style. The AI-generated text captured some elements of the journalist's scepticism and structure, but also made stylistic choices they would not have made, including clumsy constructions and passages that lacked logical coherence. Fulcrum's service can handle various writing tasks from literary fiction to casual conversation, with the company suggesting examples ranging from AI safety experts to 19th-century poets.
- Fulcrum's Echo AI generates text matching any writer's distinctive style
- A journalist tested it with their own work and found mixed but capable results
- Raises ethical questions about imitating personal writing voices
Both sides, in good faith
The strongest fair case each way — we don't pick a winner.
The case for
This technology represents a natural evolution of stylistic tools that have long existed, from writing guides to pastiche exercises. The test results demonstrate the system's current limitations—producing clumsy constructions and logical gaps that reveal its artificial nature. Used with consent and proper attribution, such systems can serve legitimate creative, educational, and productivity purposes, whether for exploring new writing approaches or assisting journalists in meeting tight deadlines. Responsibility for misuse lies with the user, not with the tool itself; restricting beneficial technologies based on hypothetical harms risks forgoing genuine innovations.
The case against
Whilst stylistic imitation is not new, the ease and convincingness of AI-generated text that mimics individuals' distinctive voices creates unprecedented risks. Without robust safeguards and explicit consent, such systems enable impersonation and creation of false attributed content far more readily than traditional forgery. The ethical concern is not whether the current version is perfect, but that the trajectory of AI improvement means future iterations will be increasingly indistinguishable from authentic writing. There is a meaningful distinction between general guidance on style and systems specifically trained to replicate an individual's voice—the latter raises fundamental questions about authenticity, attribution rights, and the potential for harmful misuse.