Artificial intelligence is advancing rapidly, and two major players, Google and OpenAI, have recently unveiled their latest AI assistants. OpenAI introduced GPT-4o, while Google introduced Gemini Live, while both of them focus on groundbreaking features like real-time conversation, live video analysis, and instant translation.
OpenAI's GPT-4o
- Features: GPT-4o can engage in real-time conversations with a response delay as low as 320 milliseconds, mimicking human-like interaction.
It can interpret easily images from your smartphone camera, assist with coding, translate text, summarize information, and generate images, fonts, and 3D models. - Access: GPT-4o's text and vision features will be gradually rolled out on the web interface and GPT app, with voice capabilities to follow.
Developers can access text and vision features in the API, with voice mode initially available to a select group. - Cost: Free to use, with caps on usage. Paid plans starting at $20 per month offer increased capacity.
Google's Gemini Live
- Introduction: Gemini Live is Google's real-time conversation AI, with plans to integrate live video communication in the future. It aims to be a versatile assistant for tasks such as job interview preparation and speech rehearsal.
- Access: Available in the coming months through Google's premium AI plan, Gemini Advanced.
- Cost: Two-month free trial, then $20 per month for Gemini Advanced.
Project Astra (Google)
- Introduction : Project Astra is an AI agent project showcased at Google's I/O conference, promising a comprehensive assistant accessible through various devices.
- Release Date: Expected later this year, offering advanced AI capabilities.
AI Safety Initiatives
- Google has invested over $100 million in AI safety research initiatives, highlighting its commitment to safe AI development.
- Sundar Pichai, Google's CEO, advocates for international collaboration on AI safety, emphasizing the importance of responsible AI development.
- Google has a team of over 1,000 engineers and researchers dedicated to mitigating bias in AI development, further ensuring the safety and fairness of its AI models.
Comparison and Safety
- While the full capabilities of these AI assistants are yet to be seen, OpenAI's GPT-4o appears to excel in audio capabilities, with realistic voices and conversational flow. On the other hand, Project Astra showcases advanced visual capabilities.
- Both OpenAI and Google emphasize safety, with GPT-4o evaluated by experts in misinformation and social psychology, and Gemini undergoing comprehensive safety evaluations, including bias and toxicity.
As these AI assistants become more integrated into our lives, it's crucial to remain risky about their safety and ethical use. While the future of AI is exciting, responsible development and oversight are paramount.