Identification of AI generated content – TOPIC 03 Sub 03 2024: Difference between revisions
(Key Participants) |
m (Re-show commented out information) |
||
| (4 intermediate revisions by the same user not shown) | |||
| Line 26: | Line 26: | ||
== Format == | == Format == | ||
To ensure | To ensure a dynamic and original format of the session, we will start with a demonstration of two tools to detect AI-generated content. First, we will show an example of how AI-generated text can be rewritten so that it evades detectors ''(AI->Human)'', using a homoglyph-based attack. Second, we will show an example with images, modifying a human image to make detectors classify it as AI ''(Human->AI)''. | ||
The demonstrations will help to illustrate the difficulties on the detection of AI-generated content, the possible negative effects of misattribution, and elicit the debate on the opportunities that arise. | |||
Then, key participants will share 3-minute insights on the topic from their fields of expertise. This would lead to a further, more informed debate between the audience and key participants, including an exchange of comments and Q&A. | |||
We will close the session by going back to the key participants and audience to see whether views have changed, by asking them to comment on the same points that they touched at the start of the session. | We will close the session by going back to the key participants and audience to see whether views have changed, by asking them to comment on the same points that they touched at the start of the session. | ||
| Line 46: | Line 48: | ||
'''Focal Point''' | '''Focal Point''' | ||
* [https://acmc-website.web.app Aldan Creo] | * [https://acmc-website.web.app Aldan Creo] | ||
Focal Points take over the responsibility and lead of the session organisation. They work in close cooperation with the respective Programme Committee member(s) and the EuroDIG Secretariat and are kindly requested to follow [https://www.eurodig.org/get-involved/planning-process/#tab-organising-a-session EuroDIG’s session principles]. | Focal Points take over the responsibility and lead of the session organisation. They work in close cooperation with the respective Programme Committee member(s) and the EuroDIG Secretariat and are kindly requested to follow [https://www.eurodig.org/get-involved/planning-process/#tab-organising-a-session EuroDIG’s session principles]. | ||
| Line 58: | Line 60: | ||
*Laurens Naudts | *Laurens Naudts | ||
*Mykolas Katkus | *Mykolas Katkus | ||
*Paulius Pakutinskas | |||
<!-- Key Participants are experts willing to provide their knowledge during a session – not necessarily on stage. Key Participants should contribute to the session planning process and keep statements short and punchy during the session. They will be selected and assigned by the Org Team, ensuring a stakeholder balanced dialogue also considering gender and geographical balance. --> | <!-- Key Participants are experts willing to provide their knowledge during a session – not necessarily on stage. Key Participants should contribute to the session planning process and keep statements short and punchy during the session. They will be selected and assigned by the Org Team, ensuring a stakeholder balanced dialogue also considering gender and geographical balance. --> | ||
<!-- Please provide short CV’s of the Key Participants involved in your session at the Wiki or link to another source. --> | <!-- Please provide short CV’s of the Key Participants involved in your session at the Wiki or link to another source. --> | ||
'''Moderator''' | '''Moderator''' | ||
* [https://acmc-website.web.app Aldan Creo] | * [https://acmc-website.web.app Aldan Creo] | ||
'''Remote Moderator''' | '''Remote Moderator''' | ||
Latest revision as of 04:30, 18 June 2024
19 June 2024 | 12:30 - 13:15 EEST | Auditorium
Consolidated programme 2024 overview
Proposals: #10 #15 #16 (#27) (see list of proposals)
Get involved!
You are invited to become a member of the Session Org Team by simply subscribing to the mailing list. By doing so, you agree that your name and affiliation will be published at the relevant session wiki page. Please reply to the email send to you to confirm your subscription.
To follow the current discussion on this topic, see the discussion tab on the upper left side of this page.
Artificial intelligence can (be helpful) prove beneficial in generating useful content. However, AI-generated content can also be used (where) to replace content that should be human-generated ( such as school exams) or to imitate authentic content (deep fakes). While deep fakes can be used for artistic and satirical purposes, (but) they can also contribute to (for) disinformation and defamation.
This session will discuss ways to identify AI-generated content:
- AI-based detection tools do not provide reliable identification of AI-generated content.
- Legal regulations may require AI tools to add a notice or watermark, but this can be circumvented.
- Cryptographic proof of authenticity for authentic content could help, but it is cumbersome.
Session description
The advent of AI has introduced a new era of content generation, with AI-generated content being employed in a multitude of formats, including text, images, audio, and video. This is utilized in a diverse array of fields, such as journalism, entertainment, and education. While AI-generated content can be advantageous in numerous ways, it also presents challenges like its potential for deepfakes and disinformation. Therefore, it is critical to be able to identify it.
Nevertheless, the principal alternatives proposed to date seem to be insufficient:
- The detection of AI-generated texts is unreliable, as it can be easily circumvented by means of paraphrasing.
- Applying cryptographic signatures to content to verify its authenticity would be cumbersome, and it would not guarantee that the content is not AI-generated. For example, if a person takes a photograph of a deep fake, the cryptographic signature would still verify the content as authentic.
- Watermarking the content with a notice that it was generated by AI can be circumvented, even if the watermark is not directly visible.
In essence, the identification of AI-generated content is a complex problem that demands a multi-faceted approach. This session will examine the challenges and opportunities in detecting AI-generated content and discuss potential ways to protect the authenticity and reliability of information.
Format
To ensure a dynamic and original format of the session, we will start with a demonstration of two tools to detect AI-generated content. First, we will show an example of how AI-generated text can be rewritten so that it evades detectors (AI->Human), using a homoglyph-based attack. Second, we will show an example with images, modifying a human image to make detectors classify it as AI (Human->AI).
The demonstrations will help to illustrate the difficulties on the detection of AI-generated content, the possible negative effects of misattribution, and elicit the debate on the opportunities that arise.
Then, key participants will share 3-minute insights on the topic from their fields of expertise. This would lead to a further, more informed debate between the audience and key participants, including an exchange of comments and Q&A.
We will close the session by going back to the key participants and audience to see whether views have changed, by asking them to comment on the same points that they touched at the start of the session.
People
Programme Committee member(s)
- Desara Dushi
- Jörn Erbguth
- Minda Moreira
The Programme Committee supports the programme planning process throughout the year and works closely with the Secretariat. Members of the committee give advice on the topics, cluster the proposals and assist session organisers in their work. They also ensure that session principles are followed and monitor the complete programme to avoid repetition.
Focal Point
Focal Points take over the responsibility and lead of the session organisation. They work in close cooperation with the respective Programme Committee member(s) and the EuroDIG Secretariat and are kindly requested to follow EuroDIG’s session principles.
Organising Team (Org Team)
- Gianluca Diana
- Vittorio Bertola
- Rokas Danilevicius
The Org Team is a group of people shaping the session. Org Teams are open and every interested individual can become a member by simply subscribing to the mailing list.
Key Participants
- Laurens Naudts
- Mykolas Katkus
- Paulius Pakutinskas
Moderator
Remote Moderator
Trained remote moderators will be assigned on the spot by the EuroDIG secretariat to each session.
Reporter
Reporters will be assigned by the EuroDIG secretariat in cooperation with the Geneva Internet Platform. The Reporter takes notes during the session and formulates 3 (max. 5) bullet points at the end of each session that:
- are summarised on a slide and presented to the audience at the end of each session
- relate to the particular session and to European Internet governance policy
- are forward looking and propose goals and activities that can be initiated after EuroDIG (recommendations)
- are in (rough) consensus with the audience
Current discussion, conference calls, schedules and minutes
See the discussion tab on the upper left side of this page.
Messages
A short summary of the session will be provided by the Reporter.
Video record
Will be provided here after the event.
Transcript
Will be provided here after the event.