1st International Workshop on Deceptive AI @ECAI2020

30 August 2020

Santiago de Compostela, Spain

About & Registration

DeceptECAI will be held on 30 August as a fully online workshop as part of DIGITAL ECAI.

Please join us on the DeceptECAI2020 Slack channel and ask questions to the authors, speakers, and panelists.

DeceptECAI2020 paper presentations are pre-recorded. If you wish to see the pre-recorded presentations, you must register with DIGITAL ECAI, which will be held between 29th August - 8th September. Registration is free.


Workshop: 30 August 2020 (online)

Submission Deadline: 25th May 2020

Notification Deadline: 18th June 2020

Camera-Ready Deadline: 15th July


Deception and its consequences are currently being enhanced by artificial intelligence (AI). Deceptive machines seem to be transitioning from a conceptual reality into our cyber, virtual or physical realities in various forms: DeepFakes, GPT-3, troll-bots. Fortunately (and also unfortunately), AI is still being used in the form of machine learning tools in the hands of humans that reason and act deceptively. However, the continuous advancement of AI might lead to truly autonomous deceptive machines that possess the ability to reason and act dishonestly without human involvement. These kinds of deceptive machines first appear as concepts in Alan Turing’s Imitation Game.

[...] it is played with three people, a man (A), a woman(B), and an interrogator (C) who may be of either sex. The interrogator stays in a room apart from the other two...It is A’s object in the game to try and cause C to make the wrong identification. His answer might therefore be 'My hair is shingled, and the longest strands are about nine inches long...'

(Alan Turing, Computing Machinery and Intelligence, 1950)

As a future scenario example, it is fairly easy to imagine artificial agents that exploit human societies in order to extract rewards, e.g. deceiving humans into voting for an entity that the machines consider to be a necessary step into achieving an ulterior goal. It may be possible that such autonomous systems might emerge from complex social interactions which they will eventually be able to manipulate according to their will. Even the intention of generation and dissemination of fake news will cease to belong exclusively to humans, when/if these autonomous agents develop their own reasons to deceive. The distinction between deceptive AI tools and deceptive AI agents will also fade away, increasing the difficulty of determining the identity of fake news authors as well as the difficulty of preventing their dishonest actions and/or holding them accountable.

DeceptECAI aims to bring together specialists from multiple disciplines in order to try to understand the various forms of deceptive machines and to address their potential threats to society. The format of the online workshop will mainly consist of 10-15 minutes pre-recorded paper presentations of peer-reviewed papers followed by 5-10 minutes Q&As discussion. In addition to this, we have an invited talk of 30 minutes + 10 minutes Q&As, and a panel to discuss the following question: How do we hold deceptive machines accountable?