Sempre on. Ma a quale prezzo? | Ivana Bartoletti | TEDxLecco

Sempre on. Ma a quale prezzo? | Ivana Bartoletti | TEDxLecco

Always on. But at what cost? | Ivana Bartoletti | TEDxLecco

🎙 Ivana Bartoletti 👥 44.7M 📅 October 14, 2026 ⏱ 17 min 👁 5 📄 science communication 🧭 2026-10-14
Available in: English (current) Français

Keywords

AIchatbotmemoryforgettingdigital detox

Summary

In this TEDx talk, privacy expert Ivana Bartoletti reflects on the impact of AI on human behavior and emotions. She begins with a personal story about using a chatbot during a difficult period, finding it therapeutic but noticing it never forgets and keeps reinforcing her pain. This leads to a broader discussion on the design of AI systems, highlighting how they are programmed to be ‘always on’ and preserve all data, yet human memory is selective by nature. She argues that forgetting is not a defect but a form of emotional intelligence, and cites Borges’ ‘Funes the Memorious’ to illustrate the paralysis of total recall. She suggests that AI should be designed to support selective memory, similar to data expiration. Bartoletti also addresses the negative effects of constant validation on children and adolescents, citing statistics on youth using AI for emotional support. She concludes by advocating for AI that prioritizes human well-being, referencing the EU Kids Act as a step forward.

161 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk offers valuable insights by shifting the focus from what AI can do for us to what AI does to us. The personal narrative makes the argument relatable, while the conceptual references (Borges, Mayer-Schönberger) enrich the discussion. Bartoletti effectively argues that memory is a design choice and that the default of permanent data retention is problematic. The argumentation is solid, though it sometimes lacks empirical depth; the statistical reference (42% of teens) is mentioned without a precise citation. Overall, the talk encourages critical reflection on AI design choices.

Scientific Rigor, Source Quality, Title Accuracy

Bartoletti refers to well-known sources: Borges’ story, Viktor Mayer-Schönberger’s book ‘Delete’, and research by GLAAD (though she says GAD). The EU Kids Act is accurately described regarding chatbots not remembering conversations with children. The title is well-matched to the content. The talk does not present original research but synthesizes existing ideas with a personal perspective. The source list is limited; viewers would benefit from direct links to the cited works.

174 words

Title / Content Match

The title 'Always on. But at what price?' accurately reflects the core theme of constant connectivity and its psychological and societal costs, as detailed throughout the talk.

Quality & Reliability

7/10

The talk blends a personal anecdote with references to literature (Borges), research (Oxford Internet Institute, GLAAD) and the EU Kids Act. While the scientific references are not fully developed and some data lack precise sources, the argument is coherent and grounded in recognized concepts of data ethics and design.

Key Moments

Cited Sources

Contribution & Novelties

The talk contributes a human-centered perspective on AI design, emphasizing the importance of forgetting as a feature rather than a bug. It connects personal experience with philosophical and ethical concepts, urging designers to consider selective memory in AI systems.

Pour aller plus loin :

  • Right to be forgotten — Core legal concept related to data deletion and privacy, directly relevant to design choices discussed.
  • Data minimization — Principle in data protection that supports limiting retention, aligning with the talk’s argument.
  • EU AI Act — The broader regulatory framework that includes provisions like the Kids Act mentioned in the talk.

99 words

Radar Profile

The radar profile shows a balanced performance with high scores in information quality and moderate scores in technical level, indicating an accessible yet thought-provoking talk. Reliability is slightly lower due to a mix of personal anecdote and scholarly references without full citation details.

Reliability 6/10