Category: Media

Featured in “ING Sector” on Artificial Intelligence

I was interviewed for the ING Sector; a quarterly magazine for corporate Netherlands, on (generative) AI. In it, I express my cautious skepticism around enterprise generative AI applications, with the quote:

we have not yet seen the big killer application of generative AI

My 2 cts: While startups, cloud providers, and consultants have been tumbling over each other for the past 2+ years trying to find (or sell) the next big killer generative AI app, we still haven’t seen it. We can effectively leverage LLMs for a variety of natural language tasks (such as information extraction), content representation (for downstream tasks), and many other useful-but-fringe use-cases.

These are all valid, but merely incremental (and in some cases, inferior, and in many cases, much more expensive) approaches over “traditional” machine learning.

The real value of LLMs and #genAI will lie in supporting us in our daily (textual; information access-related) tasks. Roll on, slope of enlightenment!

Read the full article (in Dutch) here.

Interviewed by The Netherlands Institute for Human Rights

I was interviewed by the Netherlands Institute for Human Rights, along with Joke de Groot, director HR flex at Randstad Groep Nederland, on AI and bias in the context of algorithmic matching. Read the full article (in Dutch) here: ‘Collega’s zullen nooit volledig afgaan op wat onze algoritmes zeggen’

Algorithmic bias and bias mitigation talk at the anti-discrimination hackathon

That was fun! The Online Anti-Discrimination Hackathon ran last weekend, a hakacthon co-organized by Ministerie van OCW, Inspectie SZW, Ministerie van BZK, and Hackathon Factory which revolved around “gender discrimination in data collection and labeling for automated assessment and selection of candidates.” Having ample experience in designing and developing AI-powered candidate selection systems, and the risks of algorithmic bias, I was happy to contribute to this hackathon with a few colleagues at Randstad Groep Nederland in several ways.

still of me giving the talk in front of a blue wall

Talk on Algorithmic Bias Mitigation in Automated Recruitment

First, I gave a (virtual, pre-recorded) talk on algorithmic matching, algorithmic bias, and bias mitigation in the domain of automated recruitment. More specifically, I shared how and where we use AI and recommender systems to facilitate job and candidate matching at Randstad Groep Nederland, and more generally about the challenges of bias and the opportunities of bias mitigation. I showcased both examples of misuse of AI, which results in discriminatory systems, and examples of how AI can be used to actively reduce or mitigate bias in the recruitment process. See the recording of my talk below:

In addition, me and a colleague joined live interactive roundtable sessions during the hackathon, and we brought a panel of four subject-matter experts for one-on-one sessions with hackathon teams and participants.

Read more

Featured on the European Science-Media Hub

In the context of the upcoming workshop “AI and Journalism” Bennie Mols interviewed me for the website of the European Science-Media Hub, on how we apply AI in journalism for Het FD and BNR. Read the piece here:

Or read the more in-depth background interview here:

Een lijstje zoekresultaten is geen etalage

I wrote a response on recent coverage around the alleged ‘promotion’ of anti-vax books by bol.com, suggesting it is due to “influential, steering, algorithms.” In this piece I state it’s not an issue of algorithms, but of long-tail, obscure, content. My response was published in Het Parool, as an edited letter. Read the original, unedited piece (in Dutch) below.

Continue reading “Een lijstje zoekresultaten is geen etalage”

Featured in De Volkskrant on SMART Journalism

De Volksrant wrote published an article on Dutch robo-journalism, where I’m featured on explaining our SMART Journalism recommender system + abstractive summarization plans.

Read the article by clicking on the image below:

And read more about our SMART Journalism project in our position paper:

  • [PDF] M. Sappelli, D. M. Chu, B. Cambel, J. Nortier, and D. Graus, “Smart radio: personalized news radio,” in Proceedings of the 17th dutch-belgian information retrieval workshop, 2018, p. 27.
    [Bibtex] 
    @inproceedings{sappelli2018smart,
title={SMART Radio: Personalized News Radio},
author={Sappelli, Maya and Chu, Dung Manh and Cambel, Bahadir and Nortier, Joeri and Graus, David},
booktitle={Proceedings of the 17th Dutch-Belgian Information Retrieval Workshop},
pages={27},
year={2018}
}

Featured in BNR’s “Artificial Intelligence Podcast” on AI in Media Part I: Tech

In the latest AI Podcast, I am featured, talking with Li’ao Wang on our SMART Radio and SMART Journalism projects. Stream the episode below!

“Media is smart, but needs to get smarter. That’s what we’re working on here at the FD Mediagroep. Come and see what we do with AI to make our creators and content better.”

Stream on BNR

Stream on Spotify

Interviewed in “IP – vakblad voor informatieprofessionals”

In the magazine IP (“journal for information professionals”) I am interviewed as one of three young professionals who show that ‘traditional categories and conceptual frames need to be readjusted.’

More specifically, it describes how my multi-disciplinary background, with an academic background in media studies, professional experience in the media, with a PhD in computer science, is important in bridging the gap between ‘techies’ and ‘non-techies’, and of particular value in my current role where I work on enabling AI in media. 

Digital forensics in the real world: the Ennetcom data

In the context of a high-profile legal case (involving a bunch of data acquired from encrypted “Ennetcom” phones) I assisted lawyer Inez Weski in acquiring insights and trying to understand how digital forensic tools were used in the collection of digital evidence. I did this work in the context of my PhD research on semantic search for E-Discovery. In this post, I list some of the publications that followed from my work and the case.

De Volkskrant: “Met deze eigen zoekmachine spit de politie schatten aan digitaal bewijs door”

Hansken is the search engine developed by the Netherlands Forensic Institute, and used by the police and public prosecutors. In this article in De Volkskrant, titled “Met deze eigen zoekmachine spit de politie schatten aan digitaal bewijs door,” I answered a few questions and explained my view on the role of Hansken in the court of law and digital evidence acquisition.

Click the image for the full article

NEMO Kennislink: “Het sleepnet van Justitie”

For more information on the case and my work, there’s a more in-depth piece on my work for Weski in the following NEMO Kennislink article, which details my findings and concerns with respect to using a proprietary, continuously developed, and largely black-box tool for collecting digital forensic evidence:

Click the image for the full article

Crimesite: “Hoe het pgp-sleepnet wel (en niet) werkt (#2)”

Finally, if you still didn’t have enough, there’s a blog post on crimesite which explains a bit more on the (legal) case, and some interpretations on my report and findings; 

Click to read “Hoe het pgp-sleepnet wel (en niet) werkt (#2)”

Featured in article on ‘robo-journalism’ in the Netherlands

Stimuleringsfonds voor de Journalistiek published an article on ‘robo-journalism’, where I say something about the SMART Journalism project we are doing at FDMG, which involves personalization and summarization of newspaper articles. Read it here! (pdf). Snippet:

Door introteksten te personaliseren, kun je meer doelgroepen bedienen.’ Bij het genereren van gepersonaliseerde intro’s op basis van artikelen, komt behoorlijk wat techniek kijken, vertelt David Graus, lead data scientist van het project bij het FD. ‘In de robotjournalistiek wordt nu vooral gewerkt aan het omzetten van gestructureerde data naar teksten. Wat wij willen is teksten maken op basis van door mensen geschreven teksten. Dat is behoorlijk cutting edge. We hebben daarom ook nauwelijks voorbeelden waar we ons op kunnen baseren.’

In “Denktank” on algorithms, behavioral analysis, and personalization

My debut on national TV ;-)! Denktank is a TV show where youngsters explore and think about how (current day) technology will affect them in the future. In this episode I explain some of the mechanisms behind algorithmic personalization.

Stream the episode at NPO.nl (the part with me starts at about 05:00), or see the website of Human for more information on the episode.