Professor Lorna Woods:

I am grateful for the invitation to give evidence to the committee. A way forward is to try to identify a model which is common - perhaps not a model law, but an underlying model that can then be deployed around the world in different jurisdictions.

The difficulty that is often raised with regard to content regulation is the subjectivity of content and the fact that standards differ from state to state. To address this problem, I would suggest a regulatory system that looks at the underlying systems, one that moves the focus from the content to the mechanisms that encourage content, facilitate its sharing and distribution, and select the content that comes to people's attention.

It is my contention that the platforms have developed with a disregard for the impact on the sorts of content and the sort of content that is prioritised and widely shared. This goes back to the discussion about the business model, which is about encouraging user engagement with the aim of getting more data, which is then used as a commodity. The principle put forward by Carnegie United Kingdom Trust is that there should be a statutory duty of care as regards the systems. It is starting with an assessment of risk; identifying the consequences, be they intended or unintended; and, crucially, taking steps to mitigate. Rather than leaving it at transparency, it involves asking why something has been deployed and how it can be made better.

When Google first started, it said that it viewed data as an exhaust product of the search engine business. I now wonder whether we have reached the point where people's content is the exhaust product of the data collection business and that we should be moving to try to get a cleaner engine rather than a dirty diesel one. That is what the Carnegie project is about. I am happy to share further detail. We have done quite a bit of work on this but I do not want to weigh discussions down. I would suggest that using a model that focuses on systems is beneficial in an international context because one does not have the difficulty of agreeing about difficult content. What one is looking at isn on the whole, further up the process and it is possibly easier to find common ground there so one can then be in a position of deploying a common model that can be responded to within an individual state context. I would emphasise that this can never be a silver bullet. There will probably always be a role for moderation for take-down and in some instances, possible law enforcement engagement but this sits on top of this common model. It is when one starts looking at particular items of content that the difficulty and differences arise. That is the proposal. It involves a statutory duty of care aimed at the systems - the infrastructure itself.