Medoid AI are one of the startups in the MME programme, working on an amazing solution to empower people’s voice in social media by building novel hate-speech detection and opinion mining solutions. We had a heart to heart with the founder, Anestis Fachantidis, who talked to us about their journey and their path in MME.
Tell us about your deeptech solution. Which media challenge does it tackle and why is this important?
Hatebusters tackles the problem of moderating user comments in media and social platforms to avoid the presence of hate speech, toxicity, etc.
This results to:
– Significantly reduced costs of human moderation of content entered by their users in their media and social platforms, by automatically allowing those comments with very low toxicity/hate score to appear in the platform.
– Greater user engagement in safer media communities where hate speech and toxicity are controlled and moderated efficiently.
What is unique about your solution? What sets it apart from the competition?
We bring the power and efficiency of advanced machine learning models fighting hate speech in big social platforms (e.g. Facebook) to all media and social SMES. All these businesses that don’t have the data, and the in-house expertise to build them but still need to surround their users’ communities with love and care. An enterprise level moderation solution with the unique support of our great community of “hatebusters”, NGOs and volunteers actively using our tools to detect and take down hate.
Can you shed some light on Medoid.ai’s journey so far? What have been its biggest milestones? What challenges are you trying to overcome?
We are a team of machine learning experts, our journey began by building production-grade solutions for many different spaces. Our sensitivity around hate-speech drove us to build an open, free to use hate speech detection web platform for all hate-fighters out there. On the process of understanding the problem we realized that our models would be even more valuable if utilized “closer to the hate-speech source”, the media outlets themselves and especially these that lack the resources to build such models on their own. Our greatest challenge so far is to understand the complex ways that hate speech arises depending on the context, geography and other characteristics.
You have been part of MME for a few months now. Can you please elaborate on your progress and how have you evolved during the programme?
Thanks to the high quality mentorship process implemented by Mr. John Agnantiaris and THERMI GROUP we were able to prioritize, create and start executing a product development plan, question our assumptions and research more.
What is next for Medoid AI?
Our next steps include a multi-factor validation of the Hatebusters solution, with respect to its features, user experience and user feedback processes.
What advice would you give to a founder trying to venture in the Media sector?
Let’s take an example from the brilliant name of the MediaMotorEurope initiative itself, I would advise founders to focus on the “motor” part, i.e., the actual working processes in media. The very processes that a startup seeks to either optimize (requiring domain knowledge) or disrupt (requiring deeper clients’ needs assessment).
What would you say to a startup considering applying to MediaMotorEurope?
We would definitely suggest them to apply and get the feedback and assistance needed in startups, especially those with tech founders.
Find out more about Medoid AI on their website.
