CTGT Launches New AI Platform to Remove Bias and Hallucinations

CTGT enables enterprises to deploy AI for high-risk use cases by upgrading its platform to remove bias, hallucinations, and unwanted model features from DeepSeek and other open source AI models. The upgrade allows DeepSeek to answer 96 percent of sensitive questions, a significant improvement from the original 32 percent.

CTGT showcased its platform upgrades at VentureBeat Transform 2025 in San Francisco, where it was named the top winner of the Innovation Showcase. The company has also implemented infrastructure for an insurance provider to avoid over $1M+ in financial exposure and is assisting a major bank with email compliance.

According to McKinsey, AI hallucinations caused $67.4 billion in global losses in 2024. Deloitte data reveals that 47 percent of enterprise AI users made at least one major business decision based on hallucinated output. The problem is escalating, with newer models like ChatGPT 4.5 and DeepSeek R1 exhibiting high hallucination rates.

CTGT, founded in February 2025 with funding from prominent investors, focuses on enhancing AI model trustworthiness and efficiency. The company's platform automatically removes bias, censorship, and hallucinations in real time, enabling enterprises to deploy accurate models up to 500 times faster.

For more information, visit GlobeNewswire.