Reconstructing Pelé’s Lost Goal

ORIGINAL TITLEReconstruindo o gol perdido de Pelé

Historical research, actors and AI recreate a 1959 play without original footage.

Channel
Google DeepMind
Published
Jul 14, 2026
Video
10:52
Read
2 min
Original
Portuguese
Open on YouTube
Scroll to read and the video shrinks to the bottom-right corner. Captions follow your reading language.
Reading languageAvailable in 2 languages
Machine-translated from the original Portuguese. Turn on “Show original” to compare paragraph by paragraph.
SUMMARY

In 30 seconds

Google DeepMind and the managers of Pelé’s estate present a reconstruction of the Rua Javari goal. The production combines accounts, photographs, filmed performances and generated images; the sequence is not original match footage.

  1. 1The goalThe film recalls the 1959 goal at Rua Javari.
  2. 2Incomplete sourcesA photograph and testimony do not record every movement.
  3. 3An accountAn account describes successive lifts and a header.
  4. 4Historical researchThe team reconstructed the setting with research and generated images.
  5. 5FilmingActors filmed movements on the pitch with period objects.
  6. 6The recreationThe final sequence is an interpretation, not original footage.

Takeaways

  1. The film recalls the 1959 goal at Rua Javari.
  2. A photograph and testimony do not record every movement.
  3. An account describes successive lifts and a header.
  4. The team reconstructed the setting with research and generated images.
  5. Actors filmed movements on the pitch with period objects.
  6. The final sequence is an interpretation, not original footage.
FULL TEXT

Read the edited text by chapter

Transcript only
01

A goal without footage

Memories and available records.

This film presents a reconstruction of Pelé’s goal against Juventus on August 2, 1959. Google DeepMind made the project with the managers of Pelé’s estate. The images of the play are a recreation, not original match footage.

Participants recall the Rua Javari goal as a play Pelé himself considered especially beautiful. According to the film, no television footage of the play exists.

A photograph records the header; accounts and other images help provide context. Details not visually recorded must be interpreted, and testimony cannot guarantee every detail.

One account describes the ball reaching Pelé, who lifts it over successive opponents before finishing with a header. The film presents this sequence as a reconstruction based on memories and research.

02

Research and production

Setting, actors and AI models.

The team assembled historical research to recreate the 1959 setting, including the stadium, clothing and local context. AI models were used to generate images from historical references.

Actors and a local crew filmed movements on the original pitch using period uniforms and a vintage ball. The film describes a model that extracts motion representations to help transfer movement to generated imagery.

The filmmakers present the project as a tribute and discuss their responsibility to respect Pelé’s legacy. Even with careful research, the images add creative choices to the available fragments.

03

The recreation

An interpretation presented to viewers.

The film shows the recreation alongside narration and recollections of the rainy day, the crowd and the goal. These scenes illustrate an interpretation of the play, not visual proof of its exact movements.

The production says the recreation is part of the Museu Pelé collection in Santos, intended to introduce the story to new generations.

More from this channel

5 slides38:10EN
Google DeepMind· Oct 1, 2026From deepfakes to DNA: how SynthID watermarks AIHannah Fry, Pushmeet Kohli, Jeremy Ratcliffe · Google DeepMindHannah Fry sits down with Pushmeet Kohli and Jeremy Ratcliffe to explain SynthID: a mathematical watermark now used on text, images and video, and being extended to protein sequences so DNA synthesis companies can tell AI-designed molecules from nature.BiosecurityComputational BiologyGenerative media+2
44:25EN
Google DeepMind· Sep 9, 2026How AI Is Changing Weather ForecastingHannah Fry, Peter Battaglia · Google DeepMindHannah Fry and Peter Battaglia discuss hurricane forecasts, uncertainty and WeatherNext 3.AI researchModels & training
44:41EN
Google DeepMind· Aug 26, 2026The Mathematics of AI UncertaintyZoubin GhahramaniZoubin Ghahramani explains Bayesian updating, calibration and their computational challenges.AI researchMachine Learning
53:05EN
Google DeepMind· Jul 10, 2026Inside AI Models: Neel Nanda on Interpretability and SafetyHannah Fry, Neel Nanda · Google DeepMindHannah Fry explores reasoning traces, probes and sparse autoencoders.Model InterpretabilityAI SafetyAI research
42:38EN
Google DeepMind· Jun 23, 2026When AI Agents Delegate: Security and a Possible Agent EconomyHannah Fry, Nenad Tomašev · Google DeepMindNenad Tomašev discusses specialists, oversight and correlated failures with Hannah Fry.Multi-agent systemsAgent engineeringAI security
53:48EN
Google DeepMind· Mar 10, 2026Ten Years after AlphaGo: From Go to Scientific DiscoveryHannah Fry, Thore Graepel, Pushmeet Kohli · Google DeepMindHannah Fry and DeepMind researchers revisit the match and its technical legacy.AI researchAI for scienceModels & training