Saturday, August 18, 2012

Blog update in two parts: (2) learning about Colossus

A quick thanks to all who wished me well via blog comments, emails, and calls. But this post isn’t about my recovery from that rear-end auto collision. No, this post is about something I’ve learned as a result of the accident: that there’s yet another computer system out there — one I’d never heard about before — that appears to run against the grain of my old optimistic hopes for the information revolution. This system, called Colossus, looks like a further sign that we’re seeing the rise of a new kind of computerized corporatism: what I’d like to call “cybercratic corporatism”.

I first learned of Colossus in a form letter from the insurance company of the person whose car rammed into my car. And I’ve been told that Colossus may — or may not — be used to determine an appropriate compensation for the medical part of my claim, after I file.

Colossus is evidently an expert system, loaded with data and algorithms, constantly updated and tweeked, about thousands of automobile accidents involving bodily injuries and medical claims. It was constructed in the mid 1990s, mainly at the behest of one major insurance company that wanted to cut costs, curb large payouts, and constrain its adjustors. Today dozens of insurance companies use it. It is owned by an outside computer services enterprise.

I’m not going to describe Colossus in detail here. I really don’t know a lot about it yet. I’m not even sure how central it is. But if you want to learn more and see some sources, just search the web using the following set of four terms: automobile insurance colossus computer. Doing so will yield some remarkable discussions, not to mention criticisms and allegations — e.g., about the possibilty of using it to low-ball claimants, whose “data” may or may not be entered in ways that best reflect their condition; or using it to constrain adjustors, who may receive incentives and bonuses based on their performance vis à vis Colossus.

Now, my interests are broader than Colossus by itself. Many of us — myself included — have long had hopes and expectations that the information age will favor bottom-up forces, foster greater openness, transparency, and democracy, and make hierarchies more responsive. Yet we’ve also remained warily cognizant that the information age may also be used to strengthen top-down controls, enable greater surveillance and monitoring, and facilitate secretive denial, deception, and disinformation operations. Lately there have been myriad developments in the United States and around the world that reflect all such tendencies; and which ones may prevail in the future is far from clear, far from certain.

What I’m learning about Colossus adds to my skepticism and pessimism. Colossus has a sensible economic rationality behind it, including cost-cutting and managerial rationalization. But it seems to be designed and applied as much for game theory as economic rationality. Colossus appears to be a strategic, operational, and tactical device that may be used to exert top-down control over claimants, adjustors, interactions between claimants and adjustors, and perhaps even over relations with other insurance companies.

Why is this so interesting? I have a longstanding theoretical interest in the futuristic prospects for “cyberocracy” (rule by information) and how its rise may modify and even supercede age-old “-ocracies” — like bureaucracy and democracy — not to mention various “-isms” — like corporatism and authoritarianism (see my writings, referenced below). Mostly I’ve tried to track implications for government and society at large, not so much for business, in keeping with my personal interests. Yet, I know the business world is being revolutionized by information-age dynamics, in all sorts of ways that many of you may be better informed about than I am. What’s mostly caught my attention is the development of computerized systems to suit the worlds of banking, financing, marketing, and advertising — and I’m often more dismayed than impressed by what I read. And now I can add Colossus to my sense that we are witnessing a further deeper spread of cybercratic corporatism.

[UPDATE — October 25, 2013: More information about Colossus has come to light — notably, a report by Mark Romano and J. Robert Hunter titled Low Ball: An Insider’s Look at How Some Insurers Can Manipulate Computerized Systems to Broadly Underpay Injury Claims (Consumer Federation of America, June 2012). Its Introduction observes that:
“Over the past ten to fifteen years, the payment of bodily injury claims covered by automobile or home and property insurance has evolved from a system based primarily upon the experience and knowledge of claims’ adjusters to a computer-based assessment that has the potential to be easily and broadly manipulated by insurers. This technology has enabled many insurers to increase profits by reducing the amount paid to consumers who file bodily injury liability claims, including uninsured and underinsured motorist claims. Insurers have also been able to hire less-experienced employees to handle these types of claims, since the computer programs conduct much of the decision-making. Few consumers have knowledge of these practices, while even less understand the significant impact that the practices can have on their financial lives. The authors’ primary objective in writing this report is to inform regulators about the technical complexity of this topic and the need to exercise better oversight regarding how these systems can be manipulated to the detriment of consumers. Expanding on a previous CFA report, we hope to further educate consumers filing bodily injury claims about how they can avoid unfair tactics employed by some insurers who use computer-based assessment systems and receive a fair settlement.” (p. 1)
Note the warning that computer programs may “conduct much of the decision-making.” By now, some insurance companies have apparently ceased using Colossus, while others have turned to downplay its relevance, perhaps largely because of outside pro-consumer pressures. Nonetheless, my sense is that cybercratic corporatism remains on a long-term march.]


References: Regarding the prospects for cyberocracy, see this paper and this blogpost.