18 Sep PUBLIC PENSION CORNER #57: AI in Politics and Pensions
NEWS:
I. BlackRock Shutters Two European Sustainable Funds
Citing a lack of investor interest, BlackRock has moved to liquidate two sustainable investment funds in Europe, the Impact Bond Fund and the European Sustainable Equity Fund, both of which were outperforming the markets this year:
Launched in October 2022, the Impact Bond fund aimed to give investors exposure to a dedicated, impact-focused fixed income strategy.
‘While the strategy has delivered investors exposure to this theme, the fund has not attracted the level of demand needed to achieve greater scale,’ BlackRock said.
‘After careful consideration, the directors… have concluded that the fund is unlikely to attract significant further subscriptions in the near future and believe that placing the fund into liquidation is in the best interests of shareholders.’
II. USDA Moves to End ESG Overregulation of Dairy
In response to a new report produced jointly by Heartland Institute and Consumers Defense, the US Department of Agriculture has eased ESG- aligned regulations on dairy farming:
Today, U.S. Secretary of Agriculture Brooke L. Rollins reaffirmed the Administration’s commitment to American farmers, announcing the U.S. Department of Agriculture’s (USDA) decisive action to end dairy checkoff funding for Environmental, Social, and Governance (ESG) overregulation. This action aligns research and promotion activities with the Department’s priorities and the original mission of the checkoffs. USDA remains committed to ensuring that every dollar collected from American producers serves American agriculture for the better.
“American dairy producers, cattle ranchers, and farmers pay checkoff assessments so those dollars can build demand for their products – not bankroll radical climate agendas that raise costs and constrain production,” said U.S. Secretary of Agriculture Brooke L. Rollins. “Today’s action returns the Dairy Checkoff and all research and promotion programs to their core mission: expanding markets and supporting the hardworking men and women who feed this country. We will not allow producer dollars to underwrite mandates that put American agriculture at a disadvantage.”
COMMENTARY
By Stephen R. Soukup, President and Publisher, The Political Forum
“AI in Politics and Pensions”
Like Mike Campbell’s bankruptcy in The Sun Also Rises, the emergence of AI as a potent political issue occurred in two ways: “gradually and then suddenly.” The issue has been in the background for some time – mostly manifesting in the debate over data centers. Do we want them near us? Do they raise energy costs for communities? Can they bring growth to otherwise economically troubled communities? How can we, as Americans, compete with China if data centers come to be seen as a blight on the landscape? And so on. Concerns about AI safety, however, remained mostly in the background – until last week.
Just ten days ago, Jacob Coxon, a longtime AI safety researcher, resigned from Anthropic, insisting that “Neither [OpenAI nor Anthropic] is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.” Since then, everyone and his brother, it seems, has weighed in on the question of AI safety, its political implications, and the best plan to deal (or not deal) with it. Almost overnight, the question of data centers disappeared, replaced by arguments over whether AI would kill us all and, if so, when. Given the preparedness of many of the participants in the debate, it’s clear this shift was anticipated. Still, the speed with which that happened has been bewildering.
The CEOs of OpenAI and Anthropic – nominal competitors – responded to Coxon, agreeing with his fundamental concern: we agree; we’re moving too fast. Government should slow us down. The usual suspects on the populist Left – Senators Bernie Sanders (I, VT) and Elizabeth Warren (D, MA) leapt immediately into action, joining the chorus of voices demanding a government-enforced slowdown in AI development. The usual suspects on the populist Right were split: President Trump called a slowdown a foolhardy mistake, while his former advisor, Steve Bannon, joined the Sanders crowd in demanding government intervention. Libertarian-leaning voices noted the likelihood that this was all about regulatory capture (which it probably is), and national-security hawks chimed in with a reminder about the risks of giving China an open path to global AI dominance by choosing to tie our hands behind our backs.
In short, worries about AI safety went from vague policy notions to all-consuming politics in a little over a week.
Given that I have spent the last seven or eight years arguing about the damage done by the “politicization of everything,” including capital markets, you might think that I would have a problem with this, that I would be likely to complain yet again. But that’s not the case here. This is precisely the type of issue that should be discussed and debated in the political realm. This is important, obviously. AI policy has economic implications, national security implications, religious implications, Constitutional implications, and implications for the future of humanity.
Do I wish that we had a better, more reasonable, less self-serving political class to discuss and debate these implications? Of course I do (you’d have to be nuts not to), but to paraphrase Donald Rumsfeld, you go to debate with the political class you’ve got. The fact that this has devolved to a political debate, rather than a policy debate may seem unhelpful or even destructive, but such is the nature of representative government. Policy cannot and should not be imposed from on high, based on the whims and predilections of the ruling class. Republicanism (which would have a small “r” if it didn’t come at the beginning of a sentence) demands that matters such as this be discussed, debated, and decided in public forums. If there is one overarching lesson to be learned from the Covid policy debacle, it’s that important matters must be debated openly. Public policy regarding AI safety is a matter that rightly belongs in the political realm.
Where it doesn’t belong, on the other hand, is in pension plan board meetings. And yet…
[A]n unsettling question emerged at this week’s board meeting for the California Public Employees’ Retirement System: What if the new technology leads to catastrophe? “This is humanity,” board President Theresa Taylor said Monday during a discussion with CalPERS Chief Investment Officer Stephen Gilmore. Taylor centered her remarks on the public resignation of Anthropic researcher Jacob Coxon, who wrote on social media that leading AI companies could endanger human life by the end of this decade….
Taylor, a retired state worker and former public employee union leader, urged her colleagues on the CalPERS board to consider issuing a public statement calling for new guardrails or a moratorium on AI development.
The board is obligated to prioritize the financial health of the pension fund. She argued a statement about AI concerns is appropriate because, “If we don’t have retirees, if we don’t have state workers because of AI, we don’t have a pension fund.”
I don’t want to say that Ms. Taylor’s justification for discussing the implications of AI safety is nonsensical, but it’s nonsensical. That “logic” (if that’s the right word) could be applied to any organization and any situation imaginable. “If we don’t have cows, if we don’t have hamburger eaters because the college football playoff is unfair to non-Power Five conference schools, then we don’t have McDonald’s!” It’s technically true, but…so what? The potential destruction of humanity is not really in Ms. Taylor’s purview as a pension fiduciary – just as the destructive power of college football’s playoff system is not in the purview of McDonald’s CEO (Chris Kempczinski). She may think she’s looking at “the bigger picture,” but that’s the problem – as it has been with ESG for many years.
Interestingly, the article linked above starts by noting the obvious: “California pension funds are doing well this year in no small part because of the still-booming stock market and stratospheric valuations of artificial intelligence companies.” If Ms. Taylor wanted to discuss the investment rationale for committing capital to the tech sector, that would be within her purview. If she wanted to discuss the investment rationale for subscribing to the looming IPOs of Anthropic and OpenAI, that too would be in her purview. But that would require making the investment case against those companies. Calling for a blanket government policy (which would hurt or help all investors equally) is much easier, of course. But it’s not her call to make.
AI safety is a political issue. And it should stay that way.
