Probably, an AI startup, has raised $9 million to build artificial intelligence systems that do not hallucinate. The funding round addresses a persistent flaw in large language models: generating confident but false information.
The $9 Million Bet on Reliable AI
Hallucination—when AI models invent facts, citations, or entire narratives—has become a central challenge for the industry. Probably's pitch is straightforward: develop systems that can be trusted to stick to the truth. The $9 million will go toward research and engineering to achieve that goal.
What the Funds Will Do
The company plans to use the investment to accelerate development of its non-hallucinating AI. It has not disclosed details about its technical approach, the identity of investors, or a timeline for a commercial product.
The problem of hallucination is not new. But as AI moves into regulated fields like medicine, law, and finance, the demand for reliable output has grown. Probably is one of several startups trying to solve the reliability question.
Probably's approach is focused on eliminating hallucination entirely, rather than just mitigating it. The company has not shared technical details, but the funding suggests confidence in its method.
The startup has not announced a product launch date. For now, the $9 million gives it runway to keep working on the hallucination problem that has become a bottleneck for AI adoption.
