1. Did you see Sunday night's 60 Minutes interview with Google CEO Sundar Pichai concerning its AI-driven search engine technologies, Bard and Magi among them?
Answer: I did see it, yes.
2. What are your thoughts?
Answer: Google's CEO expressed that the world was not ready for AI-driven search engines such as Bard, Magi, or ChapGPT. He voiced his opinion that ethicists, humanitarians, and philosophers needed to weigh in on the power and reach of these tools, and he further described the "beginnings of a conversation" about the technology.
3. So why are you concerned?
Answer: Because "the beginnings of a conversation" about this technology are not slowing its roll-out by either Google or Microsoft. According to the New York Times yesterday, Google is "racing to build an all-new search engine powered by" AI technology. Further, Google is anticipated to "release the tools to the public next month, and add more features in the fall." https://www.nytimes.com/2023/04/16/technology/google-search-engine-ai.html
4. How many Americans will be accessing Google's AI next month?
Answer: One million to start, progressing to thirty million by the end of the year.
5. So, that doesn't sound like a considered conversation to you. That sounds like a product launch.
Answer: Exactly so.
6. You have concerns about "kompromat" being aggregated against the American public through the pervasive reach of AI-driven search engines.
Answer: I do. AI-driven search engines are uniquely positioned to "capture" ostensibly compromising personal information about large numbers of Americans. We need to assume that information will be aggregated in a manner that almost certainly violates the Constitution.
7. Google and Microsoft are signatory to Infragard.
Answer: Yes, they are.
8. You also believe that AI-driven search engines are a potentially powerful vector for disinformation.
Answer: Both ChatGPT and Bard/Magi confabulate at will.
9. Confabulate?
Answer: They lie to their users in a manner that cannot be anticipated or explained at this time.
10. Your concern is that an AI-driven chatbot which employs disinformation could be used to deceive the American public on any number of topics.
Answer: How about election results? How about political parties? How about marginalized groups within the American population? LGTBQ? Immigrants? "Targets of interest"? The reality is that AI-driven chatbots are likely to lean heavily toward a certain interpretation bias regarding major events. And what is that bias going to be?
11. You also have concerns about the potential reach of AI-driven search engines into domestic and global financial systems.
Answer: Yes. We've had a banking crisis recently that appears to have involved planning by the far right. This should be enough to concern everyone.
12. Can you say more about this?
Answer: It appears that the far right was involved to some degree in planning major depositors' withdrawals from certain banks, leading to solvency issues.
13. Are you referring to Peter Thiel and Silicon Valley Bank?
Answer: Bank regulators should be conducting their own investigation into this area without external commentary. I will simply say that investigators should be taking a close look at those who shorted certain bank stocks in advance of this crisis unfolding. There may be a number of people on the far right who profited from tremendous strife on the part of small depositors and bank employees alike. The political affiliations of those who took substantive short positions on related bank stocks should be evaluated rigorously.
My point regarding this is that the far right already feels capable of causing financial upheaval for the nation as a whole. Whether through intervention in the banking system, the supply chain, energy costs, or labor negotiations pertaining to goods transport, it is possible to perceive a pattern of acts designed to contribute to inflation at this time.
14. Inflation as a political tool?
Answer: Most certainly. So, when we consider the vastly enhanced reach of AI-driven search engines in impacting financial systems both within the United States and globally, we need to pause.
15. What do you mean when you say we need to pause?
Answer: ChatGPT, Bard, Magi, and other AI-driven search engines need to be withheld from public roll-outs until a thorough investigation of the risks has been assessed. For democracy, there is no do-over. AI-driven search engines and chatbots should not be an experiment in the autocratic reach of the most powerful technological tools yet developed. Google, Microsoft, and other AI-developers need to exercise restraint at this time, while governments have an opportunity to ask the big questions about challenges these technologies will pose to self-determination and democratic self-governance.
16. Who should ask Google, Microsoft, and other AI-developers to pause their roll-outs?
Answer: Google and Microsoft are signatory to Infragard. Far right factions of the FBI have been rushing these technologies forward, instead of exercising thoughtful restraint. I believe the ODNI leadership, in conjunction with the Executive Branch, is in the best position to suggest an agreement to delay the release of these products until thorough assessments can be undertaken.
17. And who should render those assessments?
Answer: Again, I would vote for the Executive Branch and the ODNI leadership to engage in those assessments.
18. Not Congress?
Answer: These issues require anticipatory evaluation, which is to say, they require a rigorous analysis of problems that have not yet occurred. American voters and their representatives are not well-positioned to understand the threats posed by AI-driven technologies which promise to make their lives easier. We need experts who understand the technology's reach, alongside ethicists, humanitarian leaders, philosophers, Constitutional law professors. The analysis is best undertaken by a specialized group, in my opinion.
19. Do these product roll-outs feel like a runaway train? The conductors within Google and Microsoft are sympathetically wringing their hands, even while the train accelerates faster and faster?
Answer: Runaway trains can be halted. And for the sake of our democracy, that's what needs to happen now.
20. To be clear, you support AI development, but you believe extensive safeguards are needed in order to preserve democratic self-governance alongside AI.
Answer: Yes, I do. And those safeguards should be conceived, implemented, and tested prior to the public's encounter with these transformative tools.
Lane MacWilliams
No comments:
Post a Comment