OpenAI is enlisting third-party contractors to upload real assignments from their current or past jobs, aiming to use this data to assess the performance of next-gen AI models. This initiative seeks to establish a human baseline that can be directly compared with AI capabilities. In September, OpenAI introduced an evaluation process to benchmark its AI performance against human professionals across various sectors, an essential step towards achieving Artificial General Intelligence (AGI). Contractors must provide concrete examples of their work—ranging from documents to spreadsheets—while ensuring the removal of confidential information to mitigate risks of trade secret misappropriation and violations of non-disclosure agreements. Legal experts raise concerns about the trust placed in contractors to determine confidentiality, highlighting potential legal ramifications for both the contractors and OpenAI. This project represents a significant step in advancing AI’s ability to outperform humans in economically valuable tasks, reinforcing OpenAI’s commitment to ethical AI advancement.
Source link
