The AI Daily Brief: Artificial Intelligence News and Analysis - Should the US Nationalize the Artificial Intelligence Industry?

Episode Date: August 21, 2023

That's the argument of one Politico columnist. NLW examines the argument in depth. Before that on the Brief: A district court judge has said that AI art can't be copyrighted, but anticipates complications coming soon; Singapore workers adapting to AI the fastest; 170,000 books used in AI training dataset; and more. Today's Sponsor: Supermanage - AI for 1-on-1's - https://supermanage.ai/breakdown ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown Join the community: bit.ly/aibreakdown Learn more: http://breakdown.network/

Today on the AI breakdown, we're examining an argument that the U.S. should nationalize parts of the AI industry. Before that on the brief, a district court judge says that AI art cannot be copyrighted. The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown.network for more information about our YouTube, our newsletter, and our Discord. Welcome back to the AI breakdown brief, all the AI headline news you need in around five minutes. We kick off today with the latest ruling around AI. and copyright, and it affirms earlier rulings to suggest that at this stage, AI art cannot be copyrighted in the United States. The D.C. District Court judge involved in the case said that human beings are a, quote,

essential part of a valid copyright claim. Now, this is a case that has been ongoing for some time. It actually started all the way back in 2019 when Stephen Thaler tried to copyright a picture on behalf of an algorithm that he had created, which he called the creativity machine. At that time, back in the pre-COVID stone ages, the board basically said that there wasn't an element of human authorship, which was a prerequisite of copyright protection. Now, that decision was held up in 2022, leading ultimately to Thaler suing the U.S. Copyright Office. Thaler's claim was that the denial was arbitrary, capricious, and not in accordance with the law, but the judge disagreed. In her decision, Judge Howell wrote that copyright has never been granted to work that was, quote,

absent any guiding human hand, and which has been borne out in other cases as well. At the same time, who think that copyright should be applied to AI creations, the judge did acknowledge that things are changing fast. She wrote that humanity is, quote, approaching new frontiers in copyright, and that this will create, quote, challenging questions regarding how much human input is necessary. Now, the Copyright Office is itself trying to sort through some of these questions. In March, for example, it affirmed its previous policy that most works generated by AI could not be copyrighted, but said that in certain instances, AI-assisted materials would qualify for protection. As the Hollywood Reporter put it, an application for a work created with the help of AI can support a copyright claim if a human selected or arranged it in a sufficiently creative way that the

resulting work constitutes an original work of authorship. Now, if this feels to you like it's very blurry, you are not wrong. And as always is the case, even if no one else makes out, the lawyers will do well in this new world transition. Next up, speaking of copyright cases, one of the groups that is a little upset about AI and copyright issues is authors. Sarah Silverman, very notably, has an open lawsuit against Google and Open AI around improper training of their models on her work. And then, of course, just last week, we told the story of Prosecraft, which was a book analysis company that was forced to shut down after

backlash from authors. The latest part of this saga is a new story from the Atlantic, in which the journalist claims to have seen training data for Meta's Lama that includes more than 170,000 books from authors including Stephen King, James Patterson, Michael Poll. and others. The author writes, These books are part of a data set called Books 3, and its use has not been limited to Lama. Books 3 was also used to train Bloomberg's Bloomberg's Bloomberg GPT, Eluther AI's GPTJ, a popular open source model, and likely other generative AI programs now embedded in websites across the internet. Ameda spokesperson declined to comment on the company's use

of Books 3. Bloomberg did not respond to emails requesting comment, and Stella Bitterman, Euler AI's executive director, did not dispute that the company used Books 3 in GPTJ's training data. The author says that of the 170,000 titles, about a third are fiction and two-thirds nonfiction. Many of the titles are from big publishers, including 30,000 from Penguin Random House, 14,000 from Harper Collins, 7,000 from McMillan, and 1800 from Oxford University Press. Now, in addition to just these questions of whether books can be used to train AI models legally, there's also the question of concentration of power in AI. Books three has an independent developer behind it, Sean Presser. In conversation with the Atlantic author,

Presser said that while he's sympathetic to author's concerns, he thinks the bigger danger is a monopoly on generative AI by big corporations. He said, quote, it would be better if it wasn't necessary to have something like Books 3, but the alternative is that without Books 3, only OpenAI can do what they're doing. Now, ultimately, obviously, it will end up in front of a judge, and then an appeals court, and then probably the Supreme Court, around whether this type of material can be used to train AI, whether it constitutes fair use or not. But for now, the battle is going to be fought on the pages of magazines, on Twitter, and in the Court of Public Opinion. Speaking of the Court of Public Opinion, more echoes of the conversation we started last week about whether the AI hype cycle is over. You'll remember I did a long episode about what I think is contributing to that sense,

and then later in the week we also had Gardner labeled generative AI at the very peak of inflated expectations. A new piece in box came out over the weekend called Is the AI Boom already over? Generative AI tools are generating less interest than just a few months ago. Now, as per new data, there's not really all that much. The author cites the same similar web statistics that showed that chat GPT usage fell in June, but also flagged a story which picked up some traction last week that had two different analytics firms showing that Bing's market share was roughly the same as it was before it started integrating AI features last

January. Now, Microsoft's comment on that data has said that the firms were underestimating the numbers, but they wouldn't share their internal data, so there's no way to confirm that. What's interesting to me is less this sort of paltry evidence, and more the meta-narrative shift that this article that this is the type of article that's now getting shared and getting traction, does in and of itself say something about the current state of the market. Now, whether that's something is that media is bored with the same narratives and so we're looking for a counter-narrative, or whether it says that people really are feeling a lethargy when it comes to the technology, I think is an open and debatable question. But the fact that we're seeing more of this type of article

does seem to me to say something. Now, if some users are getting bored with generative AI, there are still tons and tons of professionals around the world who are trying to get out ahead and adapt a new set of skills that they can deploy to get ahead themselves. According to a new LinkedIn Future of Work study, Singapore's workers are the world's fastest in adopting AI skills. The report found that the share of members adding AI skills to their profile grew 20x between January 2016 and earlier this year when the report was compiled. The next five countries with the highest rate of AI skills diffusion include Finland, Ireland, India, and Canada. Although I do have to say I wonder how much self-reporting about skills on LinkedIn is the best way to identify who's actually adapting skills versus

believing that positioning themselves as having those skills is a valuable thing. Meanwhile, in another report, this one from IBM, the company says that in their surveys, executives are estimating that 40% of their workforce will need to reskill as a result of implementing AI in automation over just the next three years. AI, they say, is ushering in the age of the augmented workforce when human-machine partnerships amplify productivity and deliver exponential business value. So, heady rhetoric on the one hand, skepticism and lethargy on the other, what do you think about the truth of where AI really is right now? Let me know in the comments. Come join us on the AI breakdown Discord. The link to that is bit.ly.ly slash AI breakdown. And thanks, as always,

for listening or watching. I'll be back soon with the main AI breakdown. Before we get into the main AI breakdown, I want to tell you about today's sponsor, Supermanage. If you work in a professional setting, you probably have some version of a one-on-one meeting, either with the people that work for you or the people that you work with. Unfortunately, all too often, those one-on-one meetings become glorified catch-up calls. Don't you wish you could jump right to the stuff that really matters? That's where Supermanage comes in. Supermanage AI magically distills your team's public Slack channels into a real-time brief on any employee,

any time. Catch up on contributions, work in progress, challenges they're facing, sentiment, everything you need to show up ready for a truly meaningful conversation. And it's completely free. Visit supermanage.a.i forward slash breakdown today to start making the most of your one-on-ones. And thanks again to Supermanage for sponsoring the AI breakdown. Welcome back to the AI breakdown. Today we are talking about an interesting evolution of the narrative discussion around AI. And that's the first emergence of the first emergence of of a call to actually nationalize it. Now, obviously, the context for this is the broader conversation around AI safety, AI regulation, AI policy. Now, the specific genesis for this particular episode

comes from an opinion piece in Politico called, There's Only One Way to Control AI, Nationalization. AI's infinite potential and infinite risk requires federal ownership. The pieces by Charles Jennings, who the Politico lists as the former CEO of an AI company partnered with Caltech and JPL, and the author of a 2019 book, Artificial Intelligence, Rise of the Lightspeed Learners. Jennings writes, The AI threat is not Hollywood-style killer robots. It's AI so fast, smart, and efficient, that their behavior becomes dangerously unpredictable. Runaway AIs could cause sudden changes in power generation, food and water supply, world financial

markets, public health, and geopolitics. There is no end to the damage AIs could do if they were to leap ahead of us and start making their own arbitrary decisions, perhaps with nudges from bad actors trying to use AI against us. yet AI risk is only half the story. My years of work in AI have convinced me that a huge AI dividend awaits if we can somehow muster the political will to align AI with humanity's best interests. With so much at stake, it's time we in the United States got serious about AI policy. We need garden variety federal regulations, sure, but also new models of AI leadership

and governance. And we need to consider an idea that would have been unthinkable a year ago. We need to nationalize key parts of AI. Now, the next part of the article is spent rejecting effectively the effectiveness or likely effectiveness of the standard approaches to policy that people might be thinking about. Jennings writes, Washington is finally waking up to the importance of AI with a growing bipartisan movement advocating regulation. The meme in Congress is that we need transparency and safeguards to channel the best of AI while thwarting its most dangerous threats. If Congress simply requires that all AIs be transparent and have safeguards, the thinking goes, everything will be fine. But transparency in AI is overrated. Enacts whatever laws you like,

throw tons of money at AI transparency regulation, and we still won't have any idea how a specific AI works. I'm all for corporations disclosing data collection and AI use practices, but technical AI transparency is a mirage. Now, he says that Congressually mandated AI safeguards are more nuanced, but that ultimately they're very unlikely to be able to keep up with the speed at which AI's develop. Jennings writes, given that no one understands precisely how AI works, what guardrails could have bicker in Congress construct to protect us. How could its laws and regulations change fast enough to keep up with AI. So what is the alternative? Well, Jennings points out, America is in the lead right now because our big labs are in the lead. However, our big labs are exactly that. Big labs.

AI, he writes, is not the kind of tech that can be invented in a Harvard dorm or a startup garage in Silicon Valley. Open AI spends half a billion dollars on Nvidia infrastructure for each new AI model it launches. It has taken years of scientific study, lab research and application development, not to mention a massive investment of government dollars to construct the AI Foundation Big Tech now controls. We need a new governing body for AI in America, one that could wield the powers of the state to steer the technology toward a human mitzvah rather than a human disaster. Call it the Humane AI Commission. Now, the model that he advocates for is a familiar analogy. He wants something like the Atomic Energy Commission from 1947. He writes, The AEC model is not a perfect fit for AI. It was too slow and static, for one thing, but it is

instructive. AEC took ownership of all nuclear reactors, putting it in a position of ultimate control. The federal government's role in nationalizing nuclear weapons was that of owner, not operator. It outsourced most of the work. The military possessed finished bombs, Westinghouse built and operated nuclear energy plants, but the AEC controlled the core and had all the leverage. The AAC also owned and operated the best nuclear research labs on the planet, including Los Salamis, Oakridge, and Livermore. Historically and legally, the Atomic Energy Commission provides a useful precedent for when America creates technology that could potentially end life as we know it, a category into which AI clearly falls. Now, Jennings goes on to acknowledge that a federal

program resting control of AI's nuclear reactors from Microsoft Google at all would be a monumental and painful undertaking. However, he says all our other options are worse. Jennings argues that the pause position is viable, specifically because of China. Writing as someone who lived in China for a few years, I fear the only thing worse than a world controlled by runaway AI would be a global AI infrastructure run by President Xi Jinping. A second option is the free market option, but Jennings argues that, quote, allowing big tech to operate AI unfettered would be like Truman in trusting nuclear bombs to Westinghouse. A third option, he says, is regulation of AI by current agencies of the U.S. government. He says that narrow rules can apply, but that it doesn't actually cover the full scale of need.

So Jennings argues, what remains is the Truman option, a bold stroke of executive leadership. He writes, Within the first 100 days after the 2024 inauguration, the president announces a new national AI emergency plan. The president explains that the goal of this plan is global AI leadership for generations to come. Benevolent, peaceful leadership. Leadership that guides AI's rise as a boon to humanity. Leadership that defends the U.S. against bad actors using AI. And that installs human controls in the DNA of the most powerful AIs. Yes, that will require the federal government to take control of certain critical domestic AI resources, just as FDR temporarily nationalized parts of General Motors, Kaiser Shipyards, and other manufacturing giants to fuel America's victory in World War II.

The new Humane AI Commission will be run by a diverse team of AI experts and strive to be as apolitical as possible. Fortunately, AI policy in America has not yet been hyper-politicized. Republicans want a strong U.S. AI policy vis-à-vis China, and Democrats want racially unbiased AIs that fight climate change and create new jobs. Both agendas can be served without contradiction by an aggressive, capable new national AI plan. The only entity on Earth with both the resources and values necessary to harness AI effectively and humanely is the government of the United States. Managing AI on a global scale could well be America's greatest scientific and diplomatic challenge ever. The Manhattan Project, cubed. Now, I wanted to present this because I think it represents

a ratcheting up of the discussion of what the government should do vis-a-a-I. Of course, this is only one person's opinion. It would be incorrect to identify it as a trend in conversation that's starting to emerge. However, Politico is at this point a major publication, particularly among the DC Beltway set. It is highly likely that many of the policymakers or their staffers, at least, who are thinking about AI policy, are going to see this. Now, in addition to this just being a potentially influential argument in Washington, it also hits directly at the fears of those who are concerned about the growing rhetoric around AI safety, that the response to the fear of AI will constitute a reason for governments to seize more power because, simply put,

people are scared of it. Nationalization of a major tech industry is an example of the type power creep that those people have warned of. Now, the partisan question is an interesting one. I agree to some extent that there aren't clear, hard, and fast lines when it comes to Republicans and Democrats and their relationship to AI. But they're certainly bringing old fights into it. The discourse from the left on AI is all too often just about bias. The discourse from the right on AI is all too often just about concerns of wokeness. Certainly that seems to be where the campaign messaging around AI is, not another major new opportunity and challenge for Americans, but just another front in the culture war.

It does seem likely, however, that this is a conversation that's going to increase, not decrease. Over the weekend, the Wall Street Journal published a piece called China-U.S. Test-intelligent drone swarms and race for military AI dominance. The article was another example of where the metaphorical AI arms race becomes an actual arms race, and to the extent that that continues, and a big part of the development of AI is around its military applications, it is pretty inevitable that the U.S. government will get more, not less involved. Ultimately, I wanted to present this because my guess is that it's a discourse that's going to come up a lot more. And I think starting to understand the argument, whether you're inclined to agree or inclined to disagree vehemently, is a valuable thing to spend some time on.

Anyways, hope it was useful. And if it was, make sure you hit that like or subscribe button and even better share the episode with someone else. Appreciate you listening as always. Until next time, peace.