Massimo Leone — Augmented Images (ULiège)

Value of the Information & Strength of the Argument

The lecture’s value lies in its original semiotic framing of digital images, linking contemporary AI practices to ancient rhetorical traditions. Leone crafts a compelling argument that the distinction between technique and rhetoric is fragile, and that augmented images can serve to construct authority rather than to convey truth. The argumentation is well-structured: he builds a conceptual framework, then tests it against a specific case study, showing how the DeepDream project’s initial scientific intent is undermined by its aesthetic outcomes. The use of the augur metaphor is effective and thought-provoking, providing a fresh lens for examining the credibility of data-driven disciplines. However, the argument is largely theoretical and relies on a single case; some claims could be seen as overstated, and the speaker acknowledges his ‘Catholic provincial’ perspective as a form of self-deprecating irony.

Scientific Rigor, Source Quality, Title Accuracy

The scientific rigor is moderate: the talk references classical texts and semiotic theory, but the only specific source mentioned is the Inceptionism blog post, which is correctly identified. The speaker does not cite empirical studies or provide extensive literature, instead offering a conceptual analysis. The title’s adequacy is acceptable: the video title is general, but the talk’s specific focus on the augmented face is clearly stated at the beginning. The quality of sources is limited due to the absence of explicit citations, but the speaker’s authority and the logical coherence of the argument contribute to its overall reliability. The use of the Google blog as a case study is appropriate, as it is a primary source that illustrates the phenomenon under discussion.

The video title refers to the speaker and the symposium theme, while the actual talk focuses on the augmented face and its semiotic implications. The title is adequate but not highly specific.

8/10

The talk demonstrates rigorous theoretical argumentation, grounded in semiotics and classical rhetoric, with a well-chosen case study. The speaker is a renowned scholar, but the content is primarily interpretive and does not present empirical data.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 00:30Introduction of Massimo Leone by the moderator, highlighting his academic background and ERC project.
• 03:00Leone introduces the concept of 'augmented' from Latin etymology, linking it to the augur figure and the distinction between rhetoric and technique.
• 10:00Discussion on how sign augmentation constructs authorship and authority, and the role of devices and computation in modern pseudo-technique.
• 20:00Presentation of the case study: Google's Inceptionism blog post on reverse-engineering neural networks, and the initial scientific motivations.
• 30:00Analysis of DeepDream-generated images, the shift from scientific investigation to aesthetic production, and the comparison to augur's robes.
• 38:00Concluding remarks on the implications for semiotics and the humanities, urging suspicion of pseudo-technical rhetoric in digital methods.

• Inceptionism: Going Deeper into Neural Networks — Leone explicitly refers to this blog post by Alexander Mordvintsev, Christopher Olah, and Mike Tyka (2015) as the main case study, discussing its objectives and outcomes.