3 Amendments of Petra KAMMEREVERT related to 2020/2216(INI)
Amendment 36 #
Draft opinion
Paragraph 4 a (new)
Paragraph 4 a (new)
4a. Points out that algorithmic filters which, before or during the uploading of content, carry out an automated check on files to be uploaded and, on the basis of hash values, prevent publication of impermissible content in the first place - upload filters - are virtually unmanageable in practice, must therefore be rejected on principle and in any case are not mandated by law, since freedoms are ultimately undermined by attempts to use technology to rule out infringements of the law or make them impossible;
Amendment 44 #
Draft opinion
Paragraph 5
Paragraph 5
5. Points out that AI can be an effective tool for enforcing the rules on online content, such as illegal or harmful content or fake news, through automated content filteringidentification, and can also be used to implement the ‘notice, take down and stay down’ mechanisms; stresses, however, that AI may pose challenges toa decision to remove or block content is always an encroachment on communication freedoms that has fundamental rights, in particular freedom of expression, as well as access to information, cultural diversity and media pluralismmplications, meaning that the decision itself must always be taken by a person, in particular since AI is still unable to reliably place content in discursive contexts;
Amendment 52 #
Draft opinion
Paragraph 5 a (new)
Paragraph 5 a (new)
5a. Insists that comprehensive information that is understandable to the user be provided as to when AI is used, how it works and how decisions based on it can be challenged; points out furthermore that, per se, AI-delivered results serve only as a pointer, but never as hard evidence;