"Matt Calkins argues that companies should protect their data by running their artificial-intelligence algorithms in-house ("Don't Let AI Steal Your Company's Data," op-ed, June 29). This ignores that AI works best for big data sets, so that a program trained on industrywide records will often outcompete any single company's results. When that happens, the real question is less whether sharing is desirable -- plainly it is -- than who will reap the benefits.
Mr. Calkins's anonymous executive is right to worry that proprietary vendors like Microsoft will try to keep most of the value from shared data for themselves. The good news is that "commercial open source" offers a familiar workaround. Instead of sending their data to Big Tech, companies can agree to pool it within a single joint facility where each member can access it free of charge. Enlightened antitrust authorities should run, not walk, to approve these arrangements.
Stephen M. Maurer
Berkeley, Calif." [1]
1. Firms Need Big Data More Than They Need Big Tech. Wall Street Journal, Eastern edition; New York, N.Y. [New York, N.Y]. 07 July 2023: A.14.
Komentarų nėra:
Rašyti komentarą