Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
The UK government is set to consult on a scheme that would allow AI companies to scrape content from publishers and artists unless they “opt out”, in a move that would anger the creative industry.
The decision follows months of lobbying from both sides over whether online content should be automatically included in material that AI firms can use to train their algorithms.
Big tech companies including Google owner Alphabet have argued that they should be able to mine the internet freely to train their algorithms, with companies and creators being given the opportunity to “opt out” of such arrangements.
But publishers and creatives have argued that it is unfair and impracticable to ask them to opt out of these arrangements, as it is not always possible to know which companies are trying to mine their content.
They fear such a system would create a huge administrative burden, particularly on smaller companies, and have argued instead that they should have to actively opt in for AI systems to use their content. This would allow them to formulate licensing agreements and get reimbursed for their intellectual property, they argued.
However, ministers are planning to unveil a consultation on pursuing an “opt-out” model in the coming weeks, according to two people briefed on the plans. One added that the “opt out” model was the government’s “preferred outcome”.
The timing could slip, given that publication of the government’s much-touted “AI Action Plan” is on hold until after the October 30 budget. Others said that a growing backlash from the industry over the move might still derail the plans.
The EU has adopted a similar “opt-out” model through its AI Act, giving companies permission to mine internet content so long as the owner of the copyright and related rights has not expressly denied permission.
But media executives in the UK are deeply opposed to the proposals, arguing that it will lead to the widespread theft of their copyrighted material without remuneration.
They are concerned that the government is being too easily swayed by the arguments of deep-pocketed tech companies, with one pointing to the involvement of former and current Google executives at this week’s investment summit in London.
In a lobbying document in September, Google said that “to ensure the UK can be a competitive place to develop and train AI models in the future, [the government] should enable [text and data mining] for both commercial and research purposes”.
Media executives also warn that there is no way to verify whether an AI company has scraped their sites unless given access to their training data, something made more difficult because some AI companies scrape archives of content before coming to market, while others offer different conditions for search rather than AI-generated summaries.
Last year, the Intellectual Property Office, the UK government’s agency overseeing copyright laws, consulted with AI companies and rights holders to produce guidance on text and data mining.
However, the group of industry executives convened by the IPO — which came from various arts and news organisations, including the BBC, the British Library and the Financial Times, as well as tech companies Microsoft, DeepMind and Stability — were ultimately unable to agree on a voluntary code of practice, handing responsibility back to the government to find a workable solution.
Baroness Jones of Whitchurch, minister for online safety, said this week that the government needed to “protect the rights holders”, adding that “we are trying to find a way through that is acceptable to all sides”.
She said that she thought that the answer will be “somewhere down the middle in terms of some form of partnership and understanding”.
Chris Bryant, minister for data protection and telecoms, hosted a round table with various executives to discuss the issue last week.
A government spokesperson said: “This is an area which requires thoughtful engagement, and as part of that we are determined to hear a broad range of views to help inform our approach.
“We continue to work closely with a range of stakeholders including holding recent roundtables with AI developers and representatives of the creative industries and will set out next steps as soon as possible.”
Read the full article here