EU AI act approved by European parliament
NUJ welcomes approval of the EU AI Act requiring technology platforms to adhere to regulations on the use and development of artificial intelligence.
The Act endorsed by the European parliament on 13 March recognises varying levels of risk posed by AI and classifies technologies using a ranking system with fines for infringements associated with each. “Unacceptable risk” is considered most grave and therefore prohibited under legislation, with high-risk technologies set mandatory requirements.
In addition, developers of AI will be bound by rules when placing their products into the EU market as legislation seeks to “promote the uptake of human centric and trustworthy artificial intelligence.”
Last year, the NUJ urged Members of the European parliament to raise concerns on the use of journalistic content to train AI models. The Act outlines that AI firms must develop a policy to comply with EU copyright law and companies must provide information on the sources used to inform the training of models.
In a joint statement, the NUJ Ethics and Equality councils described the Act as “an important step amid fast-paced AI developments.”
They said:
“Approval of this Act is crucial in ensuring necessary governance of artificial intelligence, with developers rightly sanctioned for breaches. AI firms must not be granted free reign to use copyright material of journalists who face battles against tech giants currently acting with impunity.
“AI’s use to spread misinformation and disinformation or to generate deepfakes is particularly worrying to NUJ members who fear an erosion of public trust and confidence in journalism. If passed into law, this Act will support efforts to prevent harm caused by some technologies including those that perpetuate biases impacting marginalised groups. We welcome the prohibited use of technologies to categorise individuals using biometric data including in efforts identify their race, political beliefs, trade union membership and sexual orientation.”
The sinister use of AI to create pornographic images threatening the safety of women was raised by NUJ delegates at the TUC’s Women’s conference this month. The union has called for strengthened regulatory frameworks in the UK to ensure developers end the unlawful scraping of journalistic content to inform models.
Read more on the NUJ’s AI campaign.