A G7 AI roadmap lacking IP protection is one with potholes

The G7 meeting of world leaders is a clear indication of how the leading economies see AI. With the stagnation of economic growth alongside rising costs of maintaining existing public services, artificial intelligence (AI) is being seen as a quick, technological fix, that conveniently papers over the cracks inherent in society.

There is no quick fix, but G7 leaders have issued a Statement on AI for Prosperity, which effectively calls for improving innovation and adoption of secure, responsible, and trustworthy AI that benefits people.

But as Elena Simperl, director of research at the Open Data Institute points out in a recent Computer Weekly article on the UK government’s use of its Humphrey AI toolkit, using AI to process data, research policy, or write documents requires an understanding of how these technologies work, the data they rely on, and their limitations. This, she says, is the only way workers can validate AI’s outputs.

From our own experience, when an AI mishears a phrase like “generative AI” as “genitalia”, and then uses it to create a plausible sentence, it is easy to question AI’s readiness. But G7 leaders have signed up to seize the potential of AI in the public sector to drive efficiency and better serve the public.

The Business Software Alliance (BSA) praises the roadmap, describing it as “a welcome focus for G7 leaders”. The BSA says it has prioritised AI adoption in economies around the world. For an organisation that protects the intellectual property (IP) of its members, the BSA is fully committed to AI and say: “Our members look forward to working with G7 leaders to help operationalise the AI Adoption Roadmap and support the adoption of trustworthy AI across the world’s leading economies.”

Consider the irony: the software firms – some of whom are BSA members – developing AI models, are using the intellectual property of everyone who posts material on publicly accessible websites to train their models. The BSA May 30th positioning paper on artificial intelligence and copyright policy has as at the top of its wishlist for policymakers, “The development of statutory exceptions or judicial doctrines that permit AI training on lawfully accessed content.”

The final bullet point on the BSA’s wishlist on AI and copyright is: “Additional discussions regarding protections for artists from the wrongful dissemination of unauthorised digital replicas.” This is ahead of the BSA’s goal to protect AI-generated IP. One can see why, given the millions being ploughed into AI development.

There is no reference to “copyright” or “intellectual property” in the G7 Statement on AI for Prosperity. But as a society, we cannot allow big tech to steal human creativity to enrich their coffers.