From Marx to the French Ricardians, a number of contemporaries of first industrial revolutions predicted that the new technologies would destroy work. Even though rail and electricity have had only a transitory or staggered negative impact, the fear of mass technological unemployment has recently re-emerged. For two decades, the diffusion of Information Technologies (IT) has occurred at a rapid and persistent pace. Computer technology has the potential to reproduce most of the skills that are still performed by humans (e.g. speech recognition, Elliott (1998)). Computers or automation may replace thousands of workers both in manufacturing and tertiary activities; thus, popular authors such as Jeremy Rifkin (1995) prophesy the ‘end of work’ and advocate the development of new activities such as non-profit associations. This wisdom is embraced by workers, victims of unemployment, particularly in Europe (see OECD (1994) for evidence). Nevertheless, the economic literature on this question is recent and still limited; empirical evidence has failed to find a clear impact of IT on the level of employment.1 In fact, theoretical and empirical studies have concentrated on the possible skill bias in current technological change. They have agreed that new technologies shift labour demand to skilled employees. Thus, IT should explain the widening employment (and wage) differentials (Aghion et al., 1997).