Back-and-forth between Agcom and Google, on the incorporation of AI in the search engine and how it affects information traffic

Epistemia, the illusion of knowing only because artificial intelligence writes it down, risks compromising freedom of information and democracy. This is the alarm raised at the conference 'Epistemia, knowledge, AI and society' organised by La Sapienza University in Rome, a moment of confrontation between the world of research, politics, institutions and digital platforms.

And there was no lack of back-and-forth between Agcom and Google, on the grafting of AI into the search engine and how it affects the traffic of information sites. "On Google's AI Mode we are about to make a report to the European Commission, it is a clear case of impact on information," announced Agcom president Giacomo Lasorella, "there is a risk of compression of the freedom of information and the right of citizens to access multiple sources of information enshrined in Article 3 of the European Freedom Act. "AI Overview and AI Mode are a natural evolution of the search engine, we are improving the previous way of doing search not the access to information. I don't think that because of AI people stop reading newspapers, it would be worrying if they did,' retorted Diego Ciulli, Head of Government Affairs and Public Policy, Google in Italia.

But the illusion of knowledge generated by AI also impacts on the 'generation of public consensus' and thus also on democracy, as Anna Ascani (PD), vice-president of the Chamber of Deputies, pointed out. 'There is a lack of regulation,' she pointed out, 'the political decision-maker must understand that there is a problem and that we need safeguard instruments to ensure that consensus building takes place properly. The debate on AI also ties in with the issue of regulation of digital platforms, with Europe in the front row.

'The perception,' said Flavio Arzarello, Public Policy Manager, Economic and Regulatory policy, Italy at Meta, 'is that in the last five years in Europe, fear and mistrust have prevailed on the vision and approaches to artificial intelligence, with a tendency towards overprotection. Regulation should not be dismantled, but how to make it more compatible with innovation should be considered. The solution is the Digital omnibus', a Brussels proposal to simplify the regulatory framework on digital matters, 'an opportunity,' explained Meta's representative, 'to also rethink the AI Act and regulate use cases and not the technology itself'.

These moments of confrontation are 'a virtuous circuit between the world of research and the outside world, institutions and digital platforms', remarked Walter Quattrociocchi, head of the Centre for Data Science and Complexity for The society at La Sapienza University in Rome, which has had the merit of bringing the theme epistemia to the fore and which he hopes will become 'word of the year for Treccani'. During the conference, his team's research was presented, in-depth work on urgently topical issues such as 'ideologies and polarisation', 'social platforms and mental health', and 'political advertising on social media', to name but a few. "Epistemology is a more articulated problem than the circulation of fake news on social media," stressed Quattrociocchi, inviting us to be wary of those who make information about AI "without real knowledge, but only with likes and momentary celebrity".