Google's Gemini continues the dangerous obfuscation of Al technology

Key Takeaway:

Google's announcement of its new generative AI program, Gemini, marks a worrying trend in the tech industry towards secrecy and obfuscation of AI technology. This lack of transparency raises concerns about the potential risks of these powerful models and hinders their responsible development and deployment.

Summary:

• Google released its new generative AI program, Gemini, with little technical disclosure, following a similar move by OpenAI with its GPT-4 model.
• Both companies are withholding key details about the models, such as the number of neural network parameters and potential harms, making it difficult to assess their capabilities and risks.
• This lack of transparency is particularly concerning given the increasing power and complexity of these AI models.
• The absence of model cards, a standard disclosure tool for AI, is a major red flag, especially considering Google's role in their development.
• This secrecy raises ethical concerns about the oversight and safety of these powerful AI models.
• The future of responsible AI development hinges on increased transparency and disclosure from tech companies.