Ein paar Impressionen der Ausstellung https://techTURK.form-f.art. Das Märchen #maerchenvon5eeigeln / #fairytaleof5seaurchins ist nun abgeschlossen. Es war interessant, wie unterschiedlich #stablediffusion auf #prompts "reagiert" hat. Fazit: Prompts auf deutsch machen wenig Sinn, um #KIBilder zu generieren, die "hübsch" sind. Um den #bias der #Datenbasis zu sehen, sind sie größartig.
#techturk #dataminig #dataart #algorithms #algorithmart #aiart

Sometimes it's less about me implementing one thing or another so much as me getting over my personal #bias regarding it relatively being a subjectively "good" solution or not and that takes #time

When was the last time you watched a clip about #politics that aimed to bring people together, embrace #intelligence, and extinguish #bias? This is Two Minutes of Reality. 😉

#indieauthor #newbook

https://youtu.be/xqEehOvWxug

"Bad data does not only produce bad outcomes. It can also help to suppress sections of society, for instance vulnerable women and minorities."

In his new book, ‘Is Artificial Intelligence Racist?’, Arshin Adib-Moghaddam explores the relationship between various forms of racism and sexism and artificial intelligence.

#ArtificialIntelligence #AI #Bias

https://realworlddatascience.net/viewpoints/posts/2023/09/06/biased-algorithms.html

NL. Govt report: Housing seekers in the rental market are less likely to be discriminated against. Despite this outcome, housing discrimination remains a persistent problem. E.g.: people with a Moroccan sounding name have less chance of being invited to see a house.
https://www.volkshuisvestingnederland.nl/actueel/nieuws/2023/09/06/aanpak-woondiscriminatie-blijft-nodig-ondanks-daling
#law #politics #netherlands #discrimination #humanrights #bias

Ya know, if you tell me, "these three people represent a diverse group because they're #Christian, #Muslim and #Jewish", I don't consider that a #diverse group

I consider that #bias a literal #biodiversity #problem

@davep Very good example of #AI #bias.

Next was an important talk by Anthony Jehn on postgraduate earnings and the #gender wage gap in Canada at Western University, showing that wage premiums are larger for women but have declined generally over time https://www.youtube.com/watch?v=G6pSCA-qEcQ (9/11) #bias #economics

Unlocking #AI's Potential Means Getting Real About Risk

#RiskManagement #DigitalTransformation #Bias #XAI

https://www.forbes.com/sites/forbestechcouncil/2023/09/01/unlocking-ais-potential-means-getting-real-about-risk/ from Forbes

Designers often have unconscious biases that affect their design decisions and outcomes. These biases can stem from personal preferences, cultural backgrounds, or social norms. To create more inclusive and ethical designs, designers need to be aware of their biases and challenge them with research, feedback, and diversity.

#Design #Bias #Ethics

https://uxdesign.cc/biases-in-design-hiding-in-plain-sight-in-a-world-full-of-visuals-6cbe64a879f2

All of this is w/in #ADF’s #constitutional rights. But forcing #lawyers to submit to #indoctrination by ADF embodies the antithesis of #FirstAmendment values…. Somehow, #Starr is exquisitely sensitive to a perceived #bias against Carter (a SW #FlightAttendant who sent #graphic #antiabortion msgs to a fellow flight attendant) —though there’s scant #evidence SW was motivated by animus toward her faith— &resolutely oblivious to the offensiveness of required reeducation training at the hands of ADF.

Next was an important talk by Jim Raymo on the employment histories, divorce rates, and economic well-being of single mothers in #Japan at the Weatherhead Center for International Affairs at Harvard University. Single mothers are arguably even more discriminated against in the labor market than those in the US, but Raymo pulls a variety of data sources together to provide one of the most comprehensive pictures yet of the state of affairs https://www.youtube.com/watch?v=tKACvj9em90 (10/11) #gender #bias

Next was a fantastic talk by Danielle Li on algorithmic hiring, #bias, and the importance of exploration at the Stanford Digital Economy Lab. Li shows the benefits that can come from jointly optimizing for exploration and exploitation of best candidate fit (although I would argue that's a poorly defined term anyway). Highly recommend https://www.youtube.com/watch?v=8OgIEwYFcZ0 (9/11) #economics #work

Hot Take: Sony pays for favor amongst main stream reviewers. Ensure its games get a favor its and to ensure Microsoft’s get a little bit lower. Some of the comments that didn’t damage Sonys games to me proves it. If starfield were multi-plat it would been given an 8 or 9 easy.
#starfield #reviews #bias

Today Meta has released a new tool called FACET to evaluate computer vision biases. Existing AI models have repeatedly shown discrimination toward women and people of colour, which some AI experts have pointed out, is a result of bias training data. TechCrunch has the details on Meta's new tool and explains why the company's history with AI may leave some tech observers cautious.

https://flip.it/0FhkrO

#Meta #AI #Bias #Discrimination

Mitigating Age Biases in Resume Screening AI Models
"…trained an #AI model & applied #bias correction techniques …to correct for biases based on race, gender, & age. We analyzed the effectiveness of these tools in mitigating different types of bias in job #hiring algorithms, explored why age may be more challenging to eliminate than other forms of bias"
https://journals.flvc.org/FLAIRS/article/view/133236/137923
Interesting paper with a generous reference to https://economicscience.net/publications/2009-more/
#Discrimination #ageism
#JobTech #LaborMarkets

Kotek, Dockum and Sun 2023.

"Gender bias and stereotypes in Large Language Models"

https://arxiv.org/abs/2308.14921

#LLMs #linguistics #AI #bias

Myths about ‘talent’ disadvantage women in science

#Inclusion #Diversity #GenderBias #Bias

https://www.futurity.org/talent-gender-sciences-2966132-2/

Just out: A new paper I published with the great Tanja Carstensen in AI & Society, in which we explore how gender is negotiated in the German media discourse on AI and the future of work.

Gendered AI: German news media discourse on the future of work
https://doi.org/10.1007/s00146-023-01747-5

Here are some of the main findings of our analysis of 178 articles from 2015 to 2021 from German newspapers and newsmagazines:
‣ AI is portrayed as something that is shaped by society. Technological determinism is almost non-existent.
‣ The ideal worker norm is renegotiated vis-a-vis gendered artifacts such as service robots. What is the future role of women in the workforce?
‣ AI being portrayed as shaped by powerful young white men developers who are responsible for AI’s risk.
‣ It is great to see that some of the feminist and scholarly critique on AI, especially on bias and discrimination, is taken up by the media. However, we argue that this is a concept to be challenged as it masks power relations in production chains that go beyond questions of diversity, awareness and individual responsibility.

Research was funded by @boeckler_de

#Gender #AI #discourse #technology #futureofwork #work #research #sociology #paper #bias #openaccess

#Judge tears apart #Republican #lawsuit alleging #bias in #Gmail #spamfilter. Gmail isn't a common carrier and is protected by #Section230, judge rules. https://arstechnica.com/tech-policy/2023/08/judge-tears-apart-republican-lawsuit-alleging-bias-in-gmail-spam-filter/