(2 days, 15 hours ago)
Lords ChamberWith respect, it is the narrow question that a number of us have raised. Training the new AI systems is entirely dependent on them being fed vast amounts of material which they can absorb, process and reshape in order to answer questions that are asked of them. That information is to all intents and purposes somebody else’s property. What will happen to resolve the barrier? At the moment, they are not paying for it but just taking it—scraping it.
Perhaps I may come in too. Specifically, how does the data protection framework change it? We have had the ICO suggesting that the current framework works perfectly well and that it is the responsibility of the scrapers to let the IP holders know, while the IP holders have not a clue that it is being scraped. It is already scraped and there is no mechanism. I think we are a little confused about what the plan is.