@positron96.bsky.social
4. who has authority in governance.
They can absolutely fail independently. Our main argument is that when several fail together, the effects compound. I like your model-selection point too. That's probably worth making even more explicit.
Thank you for your thoughts!
They can absolutely fail independently. Our main argument is that when several fail together, the effects compound. I like your model-selection point too. That's probably worth making even more explicit.
Thank you for your thoughts!
August 19, 2026 at 10:01 AM
4. who has authority in governance.
They can absolutely fail independently. Our main argument is that when several fail together, the effects compound. I like your model-selection point too. That's probably worth making even more explicit.
Thank you for your thoughts!
They can absolutely fail independently. Our main argument is that when several fail together, the effects compound. I like your model-selection point too. That's probably worth making even more explicit.
Thank you for your thoughts!
Yes, I think that distinction is important. We try to separate these as different failure points rather than treating "representation" as one thing:
1. who/what is represented in the data,
2. whose outcomes are visible in evaluation,
3. who can identify and contest failures after deployment, and
1. who/what is represented in the data,
2. whose outcomes are visible in evaluation,
3. who can identify and contest failures after deployment, and
August 19, 2026 at 10:00 AM
Yes, I think that distinction is important. We try to separate these as different failure points rather than treating "representation" as one thing:
1. who/what is represented in the data,
2. whose outcomes are visible in evaluation,
3. who can identify and contest failures after deployment, and
1. who/what is represented in the data,
2. whose outcomes are visible in evaluation,
3. who can identify and contest failures after deployment, and
August 15, 2026 at 10:46 AM
#ResponsibleAI #AlgortithmicFairness #AIGovernance #AIPolicy #ArtificialIntelligence #InclusiveAI #TrustworthyAI #MachineLearning #DecolonialAI #DataEquity #BigData #DataJustice #AlgorithmicBias #DigitalEquity #AIforGood #SociotechnicalSystems #DataGovernance #AIAccountability #TechPolicy #Algorithm
August 15, 2026 at 10:45 AM
#ResponsibleAI #AlgortithmicFairness #AIGovernance #AIPolicy #ArtificialIntelligence #InclusiveAI #TrustworthyAI #MachineLearning #DecolonialAI #DataEquity #BigData #DataJustice #AlgorithmicBias #DigitalEquity #AIforGood #SociotechnicalSystems #DataGovernance #AIAccountability #TechPolicy #Algorithm
Takeaways (continued):
* Findings highlight the potential for advanced LLMs to improve content moderation systems across platforms.
* Findings highlight the potential for advanced LLMs to improve content moderation systems across platforms.
December 16, 2025 at 2:53 AM
Takeaways (continued):
* Findings highlight the potential for advanced LLMs to improve content moderation systems across platforms.
* Findings highlight the potential for advanced LLMs to improve content moderation systems across platforms.
Takeaways:
* LLM-based approaches show high accuracy and low false positives/negatives in detecting toxic content.
* We test robustness to adversarial attacks (input perturbations, prompt injections) and observe varying resilience across models.
* LLM-based approaches show high accuracy and low false positives/negatives in detecting toxic content.
* We test robustness to adversarial attacks (input perturbations, prompt injections) and observe varying resilience across models.
December 16, 2025 at 2:53 AM
Takeaways:
* LLM-based approaches show high accuracy and low false positives/negatives in detecting toxic content.
* We test robustness to adversarial attacks (input perturbations, prompt injections) and observe varying resilience across models.
* LLM-based approaches show high accuracy and low false positives/negatives in detecting toxic content.
* We test robustness to adversarial attacks (input perturbations, prompt injections) and observe varying resilience across models.
Datasets: X posts, product reviews, and news articles.
We benchmark both existing moderation tools (OpenAI moderation model, Llama-Guard-3) and leading LLMs (GPT, Gemini, Claude, Gemma) under normal and adversarial conditions.
We benchmark both existing moderation tools (OpenAI moderation model, Llama-Guard-3) and leading LLMs (GPT, Gemini, Claude, Gemma) under normal and adversarial conditions.
December 16, 2025 at 2:48 AM
Datasets: X posts, product reviews, and news articles.
We benchmark both existing moderation tools (OpenAI moderation model, Llama-Guard-3) and leading LLMs (GPT, Gemini, Claude, Gemma) under normal and adversarial conditions.
We benchmark both existing moderation tools (OpenAI moderation model, Llama-Guard-3) and leading LLMs (GPT, Gemini, Claude, Gemma) under normal and adversarial conditions.
December 16, 2025 at 2:03 AM
#Bowel #InflammatoryBowelDisease #ComputationalBiology #QuantitativeBiology #Bioinformatics #BiomedicalInformatics
September 19, 2025 at 7:01 PM
September 19, 2025 at 6:56 PM
Hi Sean, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
August 16, 2025 at 1:30 PM
Hi Sean, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Hi Moises, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
August 16, 2025 at 12:53 AM
Hi Moises, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Hi Joaquim, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Frontiers | Shaping integrity: why generative artificial intelligence does not have to undermine education
The integration of generative artificial intelligence (GAI) in education has been met with both excitement and concern. According to a 2023 survey by the Wor...
www.frontiersin.org
August 14, 2025 at 10:41 PM
Hi Joaquim, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Hi Tracey, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Frontiers | Shaping integrity: why generative artificial intelligence does not have to undermine education
The integration of generative artificial intelligence (GAI) in education has been met with both excitement and concern. According to a 2023 survey by the Wor...
www.frontiersin.org
August 14, 2025 at 8:03 AM
Hi Tracey, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Hi Stephen, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Frontiers | Shaping integrity: why generative artificial intelligence does not have to undermine education
The integration of generative artificial intelligence (GAI) in education has been met with both excitement and concern. According to a 2023 survey by the Wor...
www.frontiersin.org
August 14, 2025 at 7:58 AM
Hi Stephen, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Hi there, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...
Frontiers | Shaping integrity: why generative artificial intelligence does not have to undermine education
The integration of generative artificial intelligence (GAI) in education has been met with both excitement and concern. According to a 2023 survey by the Wor...
www.frontiersin.org
August 14, 2025 at 7:52 AM
Hi there, my colleague and I wrote this article on why GenAI does not have to undermine education. Please help us share it by reposting to promote dialog.
www.frontiersin.org/journals/art...
www.frontiersin.org/journals/art...