EU expects you to curate content now
You have already managed GDPR in your company and thought that’s the end of the road to be fully compliant?
Well, buckle up for the next series of regulations!
History
Privacy protection has always been in the mind of European Union bureaucrats, starting with the 1995 Data Protection Directive. It was good for a time of early web 2.0 but just two decades after — in the internet seems like a millenium — we have received a beefed-up successor, the General Data Protection Regulation or GDPR. It was 2016.
With an evolution of social media and scandals like Cambridge Analytica have redefined how we look at personal data being collected. Content curation, however, was not an issue — at least not from a legal point of view. It has changed in 2022 with a new regulation: Digital Services Act, the DSA.
State of data analysis in 2022
According to regulators in Brussels (European Commission), they have two main goals — to create a safer digital space in which the fundamental rights of all users of digital services are protected and to establish a level playing field to foster innovation, growth, and competitiveness both in the European single market and globally.
In practice it aims to harmonize rules across the European Union. Currently, digital service providers in the EU have to deal with 27 different sets of national rules, regulations, and bureaucracy. While it might bring clarity, like every grand regulation, it brings headache as well, especially mentioned Digital Services Act.
What is the Digital Services Act (the DSA)?
The DSA will require online platforms to provide clear and easy ways to report illegal content, goods, and services, and place platforms under due diligence obligations.
It basically means that they have to perform at least partial checks on the content hosted on their platform. Now, these obligations do vary a bit, and the largest platforms having biggest impact on the fabric of society — mainly through the spread of information — might cause the most serious harm. For instance, they’ll be required to design their systems and algorithms to actively tackle disinformation.
Platforms will also be required to explain any censorship decisions to their users as well as providing some sort of appeal mechanism to users who disagree with the platform’s editorial decisions. You get the idea then. The main purpose of the DSA is to make online platforms more responsible for the content they host, making them more like publishers rather than just passive hosting sites.
Seems very broad but it means that the website owners have to perform at least partial checks on the content hosted on their platform, provide fast and agile curation and give ways to appeal. On top of that, at least partial explanation — where explainable algorithms are in play — will be required, so users could understand their wrongdoing or appeal for correction.
How many algorithms do I need to run my portal?
These obligations do vary a bit, and the largest platforms where the EU claims the most serious harm already occurs will end up facing the most stringent obligations. We are mostly talking about major social media platforms and e-commerce websites with pan-european reach. For instance, those operators will be required to design their systems and algorithms to actively tackle disinformation, forcing them to have robust systems of content curation.
Explainability AD 2022 — your right to ask why.
Under new rules, platforms will also be required to explain any censorship decisions to their users as well as providing an appeal mechanism to users who disagree with the platform’s editorial decisions. Regardless if decisions will be made by humans or artificial intelligence. Basically, most mass content curations will require explainable AI and this might set the true dawn of Explainable AI.
All of this to fulfill the idea of curating and monitoring content being posted. The main purpose of the DSA is to make online platforms more responsible for the content they host, making them more like publishers rather than just passive hosting sites.
Do all websites have to implement Artificial Intelligence to operate?
Lucky no, but if one deals with content ie. comments, some sort of curation and data classification might be very handy.
This might naturally drive operations costs for small businesses, on the other side it creates a great niche of cost efficient classification mechanisms.
How to develop content curation mechanisms
TL;DR: The easiest way would be to have a set of trained classifiers, tagging content accordingly.
We at MarieAI fully support both safe and high quality content. That’s why we run a mission, to help developers create and implement fast and cost effective classification models on their websites, apps and servers with no extra cost.
Marie AI — on-device ML revolution
This is where we can proudly introduce you MarieAI — easy to use platform for creating content classifiers. MarieAI not only can create classifiers for traffic-light systems but also supports the appealing process demonstrated on the image.
On-device Machine Learning
MarieAI is available to use commercially with no extra costs. Also having your own classifiers on your server you have no additional expenses from making API calls. If you decide to use in-browser (generally speaking on-device) Machine Learning, you can use classifiers provided where no extra calls even on your servers are performed, saving resources, making operations more affordable.
You can see our video pitch here:
Create your own Machine Learning scripts without knowing how
You can make all sorts of classifiers from your data but we also have a bunch of handy samples are available here: https://marieai.com/store/
