LinkedIn Launches AI Model 'Blind Taste Test' for Premium US Users

Context LinkedIn has rolled out Crosscheck, a new feature aimed at demystifying AI model capabilities for its professional user base. This tool allows users to submit a prompt and receive two distinct responses generated by different, competing AI models from companies like OpenAI, Anthropic, and Google. Users then select the superior response before the system reveals which AI model produced each answer. Crosscheck currently processes only text-based queries, with no stated limits on the number of questions users can ask, offering broad utility for professionals.

Key Facts Hari Srinivasan, a product manager at LinkedIn, describes Crosscheck as a "blind taste test" for AI models. The feature is initially available to Premium subscribers within the United States, with future expansion planned for other countries and all free users. LinkedIn will share anonymized user data with participating AI firms to evaluate model performance across diverse professional groups.

What It Means This initiative positions LinkedIn as a practical testing ground for advanced AI. For users, Crosscheck offers a transparent way to understand the strengths and weaknesses of various AI models firsthand. For AI developers, the anonymized feedback provides crucial data to refine models based on direct professional user preferences, and its expansion will broaden its impact on AI tool integration.