Thanks for your thoughtful comments Hilary.
In our project we found that the problem was often one of misperceptions. The research funders (who have a lot of influence here) were interested in the kind of wider impact you talk about, and are strongly supportive of assessing research on a wide range of merits (and definitely not on what journal published the work). But this message was not getting through to researchers, who still thought it was all about journal impact factors and economic impact. Since then the funders continue to go in the right direction (e.g. in REF2021 the definition of impact will be 'deepened and broadened') - the challenge will be communicating this to the research community.
Agree with your broader point about consideration of the negative effects of current research culture and incentives being more important than research integrity. To an outsider with almost 15 years working (as a civil society organisation) with academics in various areas of intersection around technology and responsibility, the perverse incentives and aggressive culture are quite bewildering. Align this to the intense pressure to engage society, prove impact etc etc and there is no wonder academics are stressed, stretched and leaving the area.
Furthermore, I was on the oversight group of the recently published Quantum Tech Public Dialogue and one of its findings was that participants had assumed that 'societal implications' both positive and negative, were considered as the standard by academic researchers using public money, which underpinned their high level of trust in this group. They were very surprised when researchers admitted that it wasn't.
Participants felt strongly that wider public interest should be considered as standard in the development of QTs as well as company profit. Our analysis of previous public dialogues shows also this aspect is critical to public trust in research across tech.
I was disappointed also to discover what a narrow interpretation of 'impact' academics had to focus on. I, and many others, would like to see how this could be broadened and considerations of societal benefit, implications and involvement better embedded into the rationale and incentives for research, particularly in my area of science and technology.