Saturday, March 31, 2007

Open Source- Who is the "designer" for product liability purposes?

Engadget reported yesterday that the c,mm,n (pronounced "common") was unveiled at a Dutch auto show. c,mm,n is the world's first open source car, and the technology and plans are available for hacking and tweaking. Another open source car in the offing (but nowhere near a rolling prototype) is OScar. This is a day that I have been waiting for, when open source churns out major consumer products in meatspace. Given the recent financial and crappiness quality control issues that have plagued the auto industry, open source development promises to reduce R&D costs dramatically.

A legal issue exists, however, over defective design liability. Defective design is a products liability claim that exposes the designer of a product (as well as in addition to separate claims against the manufacturer and the seller NOTE: edited for clarity) to strict liability for injuries that result from the design flaw. An obvious example would be an iron that burns the house down because it doesn't shut off automatically- the designer should have foreseen that somebody somewhere would forget to turn the iron off.

So suppose that GM jumps on the c,mm,n bandwagon and produces it using the plans for version 1.0, and several years later a number of accidents reveal a design flaw that cause the vehicle to explode in rear-end collisions when both the left turn signal and license plate frame lamp are activated. So who is liable as the designer?

The open source development model has proven to be outstanding for developing software. However, since software generally doesn't kill people (at least not yet), these issues have yet to be resolved. I can see courts resolving the issue one of three ways:

1) Adoptive design. This is a kind of due diligence argument- if the General wants to slap their name on the car, then they should have made sure the plans work. This is a fairly common sense resolution- since the car is open, GM is the designer of the car if they have the opportunity to alter the design but do not do so. Clearly, if they had modified the design there would be no question as to liability; but by building the car as is, they have endorsed the design and adopted it as their own.

This is the best solution, as it encourages manufacturers to do their homework before releasing the product, and it allocates the risk to the party best able to bear it- i.e. the party who will actually profit from the design.

2) All designers are liable. This would be the worst possible solution, in my estimation, but there will be pressure coming from proprietary industries to persuade courts to embrace this approach. Imagine the onslaught of amicus briefs from Microsoft, Apple, and the patent bar arguing that "fundamental fairness" requires open source designers to be subject to the same liabilities that they are. Couple it with the lobbying and PR advantages that these industries enjoy, and I can picture the endless commercials on CNN or the Sunday morning talking head shows touting this line.

If a court were to adopt this line, and each designer were potentially on the hook, the litigation would be virtually endless. Before we even get to the ultimate issue of liability, there will be years of collateral litigation- is this kind of suit covered under a designer's homeowner's insurance? How does a court allocate the damages? If the defective component was 99% designed by one person, but 1000 other people contributes the remaining 1%, would all designers be jointly and severally liable equally or only liable for their actual contribution? If the community is jointly and severally liable, then the designers fight amongst each other in a separate trial over what their respective contribution should vis a vis each other, which would itself require a trial on who designed what. I could go on, but I am still only in my pajamas, and thus my brain is only running at 80%.

If courts adopt this rule, it would really sound the death knell of the open source movement, which is what proprietary industries really want anyway.

3) No design defect liability. A court could simply hold that an open source product is not subject to design defect liability. This option has the benefit of simplicity, but it seems unfair to plaintiffs. Design defect and manufacturer defect claims are different tort theories, not two different ways to go after somebody's pocketbook. Theoretically, a defectively designed car could be perfectly manufactured- thus no manufacturer's liability. The plaintiff's remedy is to go after the designer.

If open source products are immune from design defect, then the driver of a perfectly built but defectively designed Chevy has a remedy available to him/her that the driver of a perfectly built but defectively designed c,mm,n lacks. A market solution exists for the driver- if the driver can't bear the risk that she might not be able to recover on a design claim for the c,mm,n, then she can buy the Chevy. But what of injuries to third parties? A passenger in the defective car would be just as injured by the explosion, but would not have an opportunity to account for the risk through market action. Should that person only hitch rides from drivers of proprietary cars? What of the other vehicle in the accident?

This rule would also encourage the manufacturers of open source products NOT to perform due diligence, because taking the time to ensure that the car is well designed puts them at risk if they miss something. Obviously, we want safe products on the market, and any rule that encourages blindness to safety ought to be avoided.

I think my first option, adoptive design, works best for all concerned. It is fair to plaintiffs and defendants; it encourages manufacturers to check the work of the community; and it guarantees that open source development will continue and expand into real world consumer products. I also think I stumbled blindly and half asleepedly into a law review article. I will keep thinking about this and maybe develop it.