On 16 January, the Chair on AI and Democracy of the Florence School of Transnational Governance (STG) hosted the "AI and Democracy Annual Lecture" featuring Professor Luciano Floridi, a renowned philosopher and the Founding Director of the Digital Ethics Center at Yale University. The lecture delved into the philosophical aspects of Artificial Intelligence (AI) and explored its impact on global democracies.
The event was introduced by Professor Daniel Innerarity, Chair in AI & Democracy at the Florence STG. Professors Floridi and Innerarity were also joined by discussants Shamira Ahmed, STG Policy Leader Fellow, and Stefania Milan, Professor of Critical Data Studies at the University of Amsterdam and Research Associate at the STG.
The lecture posed critical questions about the role of AI in democracies, examining whether it serves as a beneficial force or a potential obstruction.
Is Artificial Intelligence actually intelligent?
Professor Floridi engaged the audience in an insightful exploration of whether AI has agency.
In philosophy, the concept of agency refers to the capacity of an entity to act, make choices, and exert intentional control over its actions. Agency involves the ability to initiate and direct one's own actions, demonstrating a form of self-determination and purposeful behaviour.
The attribution of agency to AI is often debated and depends on the level of autonomy, intentionality, and consciousness associated with the system.
Floridi argued that AI systems exhibit a form of agency in specific domains, where they can carry out tasks, solve problems, or make decisions within predefined parameters. This agency is limited to the capabilities programmed or learned by the AI. At the same time, it has zero intelligence.
"I wouldn't consider a pocket calculator intelligent. I don't consider Chat GPT intelligent either," Floridi said. “With AI we experience a divorce between the ability to solve problems with success in view of a goal (agency) and the need to be intelligent to do so.” Pointing at his mobile device, he added: “This smartphone plays better chess than anyone else in this room. I bet. But it has zero intelligence.
Floridi explained that the new AI, such as ChatGPT, a chatbot developed by OpenAI based on a large language model, simply leverages statistical techniques; all it does is highlight correlations between things and events.
While ChatGPT can generate contextually appropriate responses, it is important to note that it does not have true understanding, consciousness, or awareness. It operates based on statistical associations learned from data and lacks genuine comprehension or intentionality.
AI and democracy
Addressing the ongoing academic discourse on AI, Professor Floridi assessed its intricate relationship with democracy and delved into its influence on philosophical and political discussions surrounding this revolutionary technology.
He acknowledged that AI can help mitigate the crisis of democracy, but not by replacing elected representatives in their ability to make decisions:
“Anyone who thinks they can actually fix Chat GPT to make sure that it always provides correct answers does not understand the technology in question. And I'm talking about the people in power [...]. It's a statistical tool. It's intrinsically not able to provide certainty. It provides most likely, almost certainly, but never certain.”
The event delved into the question of how we can use AI to have a positive impact on democracy. Floridi suggested we need to start to rethink democracy itself before looking to AI for help.
Drawing a parallel, he highlighted that while antibiotics are undoubtedly beneficial, their misuse can lead to resistance. Similarly, Professor Floridi suggested that viewing democracy as a panacea for global challenges might lead to its misuse. He advocated for a nuanced understanding, emphasising the need for a different and better democracy rather than merely advocating for more democracy.
At the core of the way we restructure democracy, Floridi mentioned the possibility of creating and separating digital sovereignty. This implies that AI can be used to ensure that control over essential questions is not concentrated in the hands of a few powerful entities (such as big tech companies) but is distributed more broadly, fostering a more democratic approach to decision-making.
The efforts underway through the AI Act in Europe, Floridi argued, are forging the path to effectively regulate this ever-evolving tech. European legislation, with an improved understanding of AI technology, is moving in the right direction to protect democracy without stifling innovation.
Watch Floridi’s full speech here on our YouTube channel and revisit the best moments of the lecture on our Instagram channel.