Kentaro Toyama argues that ICT4D is full of hype and light on impact for one simple reason: rare is the technology that unilaterally solves a problem. Technology needs humans, preferably well-intentioned and effective ones, to make an impact on poverty.
I agree with just about everything Toyama has to say, with one caveat and one addition.
The caveat: we can look for interventions that generate positive impact irrespective of the intentions or effectiveness of the implementing body, or interventions that implement themselves, once set in motion.
For example, technologies that disseminate price information have been shown to be valuable in reducing market failures (per the studies by Robert Jensen and Jenny Aker that Toyama mentions). These technologies hold promise perhaps for the very reason Toyama is critical of most ICT4D interventions. Specifically, they do not rely on a motivated and well-intentioned NGO, government, or firm in order to have positive effects. Self-enacting programs may be the exception to the rule, but more effort should go into identifying them.
The addition: evaluate.
Evaluation shows us whether ICT4D solutions can reduce poverty, and under what circumstances. Clearly, a technology that potential users don’t adopt is a technology that is not working. But a high rate of adoption is not sufficient evidence of success. Many widely adopted tools do not reduce poverty. Those concerned with fighting poverty—not selling gadgets and hobbies and toys—must know which ideas actually reduce poverty and which just sell, or worse yet, lead to behavior that undermines long-term health and prosperity. A new technology for smoking tobacco, which generates widespread adoption and usage and thus passes a “market” test, would not excite our charitable instincts. Heavy usage should awaken profiteers, but it does not establish a tool as effective in the arsenal against poverty.