A forthcoming article in the NYU Law Review by Jeanne C. Fromer argues that the software industry is increasingly turning to trade secret protections:
With secrecy becoming more robust for the software industry and with this shift making it so much more difficult, if not impossible, for independent discovery, reverse engineering, and knowledge transfer from departing employees, the landscape in the software and computing industry seems to be shifting. In recent years, this industry has seen increased resort to trade secrecy, as reflected in both legal scholarship and media reports. With these shifts, the industry might start experiencing a slower pace of innovation, offering fewer software choices for consumers, and enabling fewer competitors to enter the industry. As a corollary, there also might not be sufficient incentive for newcomers or competitors to innovate in the face of an initial mover’s robust trade secret.
The article analogizes current innovations in the tech industry to Willy Wonka’s Oompa-Loompas. Wonka did not enlist his peculiar workforce for their dexterity or tolerance of poor working conditions but for their aid in protecting his intellectual property. Worrying that a normal workforce might leak the recipes to his exotic treats, he conscripts a workforce that need never leave the factory.
The author argues that several recent technological innovations make trade secret protection increasingly attractive for software developers. Cloud computing can keep source codes out of reach for potential imitators. Programmers are increasingly using machine learning to develop software that is practically impossible to reverse engineer. Automation can be used to replace workers who would pick up valuable skills and transfer them to future employers.
The use of automated Oompa-Loompas to create trade secrets, rather than patents, has serious drawbacks for society at large. Trade secrets can (in theory) permanently cloister innovative ideas as the possession of a single enterprise. Consumers bear monopoly prices, just like a patent, but competitors lose the opportunity to study an innovation–the silver lining of the patent system with clear benefits to innovation. While software patents raise costs for consumers, hamper downstream innovators, and provide dubious incentives to innovate, trade secrets keep that same information out of reach.
One could argue that it is within the rights of a company to not share information for the purposes of competitive advantage. It’s a reasonable position, but it’s equally reasonable to argue that the government need not bend over backwards to protect corporate secrets. Fromer, however, finds a number of troubling policies that give trade secrets more protection than they deserve.
A body of legal precedent has accumulated determining what information employees can carry with them to their next job, what constitutes valid reverse engineering, and what contracts stipulating secrecy are actually enforceable–most of which is deferential to the firm holding the trade secret. Additionally, firms claiming trade secrets have also been spared scrutiny by antitrust regulators.
If the move toward trade secret protection for software takes off and innovation appears to be suffering, governments should consider ending, or at least scaling back, their deference to software producers on these issues.