Artificial Intelligence Perplexity AI Safety Research 2025: Key Findings & Future Implications for Ethical AI