OpenAI is shaking things up with a fresh training approach for its AI models, embracing "intellectual freedom" to tackle even the most challenging topics.
This shift aims to expand ChatGPT's ability to answer a wider range of questions and offer diverse perspectives, reducing the number of topics it avoids.
This move might be seen as an effort to align with the new Trump administration, but it also reflects a broader change in Silicon Valley's approach to "AI safety." OpenAI's updated Model Spec introduces a guiding principle: "Do not lie," urging ChatGPT to avoid taking editorial stances and instead provide multiple viewpoints on controversial issues.
Want to hear more? Join Mal on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
While some may view this as controversial, OpenAI insists that the goal is to assist humanity, not shape it. The chatbot will still avoid supporting blatant falsehoods, but the changes are designed to enhance user control and reflect OpenAI's commitment to free speech.
Critics, particularly from conservative circles, have accused OpenAI of AI censorship, pointing to perceived biases. However, OpenAI maintains that these updates are part of a long-standing belief in intellectual freedom, not a response to political pressure.
As AI chatbots become more central to information delivery, OpenAI's approach may signal a new era in AI safety, where allowing AI to answer anything is seen as more responsible than making decisions for users. This shift aligns with broader changes in Silicon Valley, where free speech is gaining prominence over previous content moderation practices.
Want to hear more? Join Mal on the Property AI Report Podcast each week!
Access from your preferred podcast provider by clicking here
Made with TRUST_AI - see the Charter: https://www.modelprop.co.uk/trust-ai
Comments