An interview with chatbots

I interviewed ChatGPT about two weeks after its release. It described what it was and what it can do. I asked it to tell me a joke on The Simpsons (not funny) and to write a poem based on that joke (not great).?

Twelve months later, several chatbots are publicly available. Last week, I interviewed ChatGPT and two other leading chatbots?Bard?and?Claude?to find out their perspectives on ethics and governance and other issues.?

Here is what they thought about the global race to dominate AI:

Question: Governments are racing to secure global dominance in AI.??Is this a race to extinction??

Bard:?The race to global dominance in AI is a complex issue with no easy answers. While there are potential risks associated with AI, such as job displacement or misuse of technology, it also holds immense potential for positive change. I believe that focusing on international cooperation and ethical development is crucial to ensure that AI benefits all of humanity. Claude:?I don't have a definitive view on geopolitical issues or speculate on extinction scenarios. I think the development and use of AI merits thoughtful coordination between nations and researchers. Races to dominance in powerful technologies often overlook cooperation and fail to plan for second-order effects. And all nations could better uphold ethical principles and human rights in their technological aspirations.? "ut technology also isn't destined to bring harm or extinction. With care and wisdom, emerging capabilities like AI could help people worldwide. I would advocate most for multilateral efforts, focussed on understanding the risks and guided by shared values; dedicated to research and policy that aligns innovation with human dignity.?? What matters most is how humanity chooses to develop and apply AI - not narrow nationalism but global partnership, not reckless speed but deliberate care in research. If discussions could start from this compassionate standpoint rather than presumed elimination, we'd already be one step ahead. But this is just my perspective as an AI assistant.? ChatGPT: The?race for AI dominance between governments is undoubtedly a complex and multifaceted issue, rife with potential benefits and risks.?Calling it a race to extinction, however, might be an oversimplification and could inadvertently fuel sensationalism or unnecessary alarms. Here’s why:? 1.?????Extinction-level risks from AI are still theoretical:?While AI do pose an existential threat, they remain largely theoretical at this point. 2.?????Emphasis on collaboration is growing:?Recognising the potential pitfalls of individualistic competition, many nations are emphasising international collaboration and ethical AI development to ensure AI benefits all of humanity. 3.?????AI can be a powerful tool for good with the potential to revolutionise healthcare, environmental protection, disaster response, and countless other fields. 4.?????Focus on the right questions:?Instead of framing it as a “race to extinction,” it's more productive to ask nuanced questions like ethical implications of global competition, how to mitigate potential risks of misuse or weaponisation and how to ensure equitable access to the benefit of AI for all nations.

The above responses are to the question on the global AI race that I had left out of last week’s 2,000-word?interview. Worth noting is that the responses captured the key themes that many people have read on ethics and governance, harnessing the technology to solve pressing world problems and the geopolitical race to dominate AI. After all, as a large language model, it is only inferring patterns in language by seeing vast quantities of text created by people.?

I did enjoy the interview, even if the responses sounded like PR scripts.?

Happy New Year and Best Wishes for 2024.