Elon Musk has stated that artificial intelligence companies have utilized all available human knowledge and are now turning to “synthetic” data to train their models.
According to the world’s wealthiest individual, technology firms will need to rely on AI-generated materials to develop and refine new systems, a practice that is already advancing rapidly.
Musk commented in a livestreamed interview on X, saying, “Human knowledge reserves are being depleted in training AI. That’s essentially what happened last year.”
AI models like ChatGPT are fed vast amounts of internet data to learn patterns and predict upcoming words in a sentence.
Musk emphasized that the only solution to the lack of training data is to shift towards synthetic data generated by AI.
Companies like Meta, Microsoft, Google, and OpenAI are already utilizing synthetic data in their AI endeavors.
However, Musk cautioned against the tendency of AI models to produce “hallucinations,” which could pose risks in using synthetic data.
During a live streaming interview with Mark Penn, Musk highlighted the challenge of distinguishing between hallucinations and real answers in the AI material processing.
Controlling high-quality data is a key issue in the legal landscape of AI development, with challenges arising regarding access to copyrighted materials for model training.
Source: www.theguardian.com
