My Observations About Artificial Intelligence
On Monday, June 17, 2024, I spoke at the AI/COMM Conference co-hosted by Jeff
Whether a leader is making a decision about future growth prospects of a business, a financial advisor is attempting to value growth in a business sector relative to an investment, or a government official is trying to assess systemic risk, the one common problem is risk assessment that results from the ripple effect of a highly concentrated, but poorly understood, threat.
Businesses have in-depth knowledge and core competency with respect to the markets in which they participate. They may have vertical market experts (e.g. oil and gas, agriculture, manufacturing), or geographic market experts (e.g. China, the EU, Latin America). Similarly, financial advisory firms and investors may have sector experts (e.g. emerging market experts, energy sector investors, healthcare experts). The federal government is organized into cabinet departments (Energy, Transportation, Commerce).
What is missing is a systematic effort to assess the interdependencies among various market sectors. Yet, the interdependencies are what take all leaders and, by extension, the public, by surprise and have profoundly disruptive effects.
In 2008, we knew at Pitney Bowes that some portion of the mail that we processed through our machines came from the financial services industry. We even had a good idea about the amount of mail that depended on solicitations for various kinds of credit, such as credit cards, home mortgages, home equity loans, and car leases.
What we did not understand was the underlying financial vulnerability of financial institutions to the explosive growth of derivatives and how they would abruptly reduce the volume of credit granted or solicited if these derivatives transactions suddenly halted or reduced significantly in volume. Mail volumes going through our machines dropped by 25% in the last quarter of 2008 and dropped another 15% in 2009.
My successor Murray Martin had to deal with what would be called a “black swan” event, that is, a high impact, exceptionally low probability catastrophe that no one could have seen coming. Moreover, because of steps taken in the financial services industry to secure the safety of financial institutions, whole chunks of mail volume disappeared and did not come back when the economy recovered.
Today, we are learning about the twin ripple effects of the rapid drop in oil and gas prices and the slowdown in the Chinese economy. As consumers, we are getting a windfall from lower gasoline prices, but a whole range of industries are suffering from the combined impact of reduced orders from China, which has cut back on capital spending and construction and is dumping excess commodities like steel into world markets, and a worldwide reduction of spending by the energy industry.
I have learned about some interdependencies of which I was not previously aware. For example, among the cutbacks in purchasing affected by oil and gas production cutbacks are the reduced purchases of helicopters that transport oil workers to remote offshore rigs. This cutback would surface in the aerospace industry, not in the oil and gas business.
The most serious and ominous interdependence was highlighted in the Sunday, November 8, 2015, issue of The New York Times Sunday Review section, in an article on cyber-threats by Kate Murphy entitled “Underground Cyberthreats.”
//www.nytimes.com/2015/11/08/sunday-review/the-cyberthreat-under-the-street.html?_r=0
A less informed member of the public might think of the Internet residing in a “cloud,” which is a concept that implies a system of wireless connections. I believed that we had a highly decentralized and dispersed data communications system, since the Defense Department created the Internet to eliminate the risk of dedicated long land lines that were built by ATT when it had a telecommunications monopoly.
Not surprisingly, I was shocked to find out from the New York Times article that there are about 80 communication nodes that carry a huge volume of Internet traffic. If any one of them were shut down, either through mechanical failure, sabotage, or terrorism, the economic effects would be devastating all over the world.
Similarly, we have only two train tunnels underneath the Hudson River, both of which need extensive repairs as a result of the damage done to them by Super Storm Sandy. Experts predict that each of them will need to be taken out of service for at least one year in the next 15 years to avoid a collapse that endangers people on the trains. If either one is shut down today, the ripple effect is such that about 75% of rail track would be shut down, not 50%, because of track and signaling configurations. Yet no effort is underway to fund the building of a 3rd tunnel.
There are many other examples of high risk-concentration with which we live in blissful ignorance. Some of them, like the ripple effect of a financial meltdown on the mailing industry, are virtually impossible to anticipate. However, shame on us for not addressing those we can anticipate and address, such as the problems with the tunnels.
One of the problems with the way we discuss our crumbling infrastructure and other major societal issues is that we discussing problems in large percentages, such as the Federal Highway Administration statement that 21.5% of our bridges or tunnels are either deficient or obsolete.
//www.fhwa.dot.gov/policy/2013cpr/overviews.cfm#2h
These kinds of numbers are important to communicate to the public, but what would be far more powerful is to identify specific points of vulnerability to the degree that we do not give a roadmap to terrorists, in order to shock the public and our elected officials into doing something. Moreover, what we really want to know is which bridges, if they were taken out of service, would have the biggest impact on the public. This kind of broad statistical analysis simply is insufficiently informative.
The Regional Plan Association, of which I am a Vice Chairman, has done exactly this throughout its history. Recently, Tom Wright, the RPA’s current CEO, did a small video in which the RPA clearly and powerfully described the risk of a shutdown of either of the cross-Hudson rail tunnels.
www.rpa.org/tunnel-trouble-crumbling-infrastructure-is-putting-region-at-risk
Over the next few years, as the RPA develops and publicizes its Fourth Regional Plan, it will be doing more of this kind of work to identify the highest priority, highest risk infrastructure issues in the Tri-State Region.
Our elected officials and candidates generally do a poor job educating the public on these risks that can give rise to “black swan” events. We need to demand more of them as we move into the 2016 Presidential campaign. I intend to use this blog and other platforms to get this message across.