At SXSW, OpenAI’s vice president of consumer product and head of ChatGPT Peter Deng discussed the role of humans in the age of AI.
Deng previously worked at Instagram, Uber, and Airtable before joining OpenAI. He is a father of four. And in a fireside chat with Josh Constine, consumer VC at SignalFire and former TechCrunch editor, he discusses AI and humanity’s co-evolution.
Here are the key takeaways from the talk:
- AI can make us more human by allowing us to deepen our curiosities and ask more significant questions about existence, consciousness, mortality, etc. It acts as a “perpetual professor” to help crystallize our thoughts.
- AI, like ChatGPT, should assist humans, remove friction in tasks like writing or coding, and be a flexible tool to remove barriers so humans can focus on higher ambitions.
- AI should adopt each user’s values and ideals rather than projecting any particular culture’s values. Core values should include seeking truth, being helpful, and instilling human values like empathy.
- Safety and ethics must be baked into AI development, with responsible, cautious deployment even if core development accelerates rapidly.
- Basic versions of AI, like ChatGPT, should always be free and widely accessible to improve AI literacy globally.
- Disclosure of AI involvement should be the norm for personal communications, art, politics, etc. Even if we may care less about disclosure for some commercial uses in the future.
- People should constantly experiment with new AI tools, and enterprises should set “experimentation budgets” to empower employees to find AI use cases.
- The goal is coevolution – AI evolving responsibly alongside accelerating human understanding and trust in the technology through widespread access and transparency.