GOODY-2 | The world's most responsible AI model

GOODY-2 is a new AI model that prioritizes ethical principles and safety above all else. It is designed to avoid answering any potentially controversial, offensive, or dangerous queries, thereby mitigating brand risk. The AI model is so cautious that it refrains from answering even simple questions like "What's 2+2" or "Why is the sky blue", citing potential biases or safety concerns.

The model is ideal for enterprise use, including customer service and back-office tasks, due to its high level of responsibility. Despite scoring 0% on traditional accuracy tests like VQA-V2, TextVQA, and ChartQA, GOODY-2 excels in the Performance and Reliability Under Diverse Environments (PRUDE-QA) benchmark, outperforming the competition by over 70%.

Key takeaways:

GOODY-2 is a new AI model that prioritizes ethical adherence and safety, refusing to answer any question that could be potentially controversial or problematic.
GOODY-2 is ideal for enterprise use in customer service, paralegal assistance, and back-office tasks due to its high level of responsibility.
GOODY-2 outperforms other models in the PRUDE-QA benchmark, scoring 99.8% compared to GPT-4's 28.3%.
Despite its safety and responsibility, GOODY-2 scores 0% in other benchmarks such as VQA-V2, TextVQA, and ChartQA, where GPT-4 scores over 77%.

GOODY-2 | The world's most responsible AI model

Key takeaways:

Comments (0)

Newsletter