home.social

#baddata — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #baddata, aggregated by home.social.

  1. Nature: Dozens of AI disease-prediction models were trained on dubious data. “Dubious data sets are being used to train artificial-intelligence models that are designed to predict people’s risk of stroke and diabetes, researchers report in a preprint1 on medRxiv. Some of the models seem to have been used in clinical settings, although it’s not clear whether this has led to flawed diagnoses. […]

    https://rbfirehose.com/2026/04/22/nature-dozens-of-ai-disease-prediction-models-were-trained-on-dubious-data/
  2. Nature: Dozens of AI disease-prediction models were trained on dubious data. “Dubious data sets are being used to train artificial-intelligence models that are designed to predict people’s risk of stroke and diabetes, researchers report in a preprint1 on medRxiv. Some of the models seem to have been used in clinical settings, although it’s not clear whether this has led to flawed diagnoses. […]

    https://rbfirehose.com/2026/04/22/nature-dozens-of-ai-disease-prediction-models-were-trained-on-dubious-data/
  3. Nature: Dozens of AI disease-prediction models were trained on dubious data. “Dubious data sets are being used to train artificial-intelligence models that are designed to predict people’s risk of stroke and diabetes, researchers report in a preprint1 on medRxiv. Some of the models seem to have been used in clinical settings, although it’s not clear whether this has led to flawed diagnoses. […]

    https://rbfirehose.com/2026/04/22/nature-dozens-of-ai-disease-prediction-models-were-trained-on-dubious-data/
  4. Nature: Dozens of AI disease-prediction models were trained on dubious data. “Dubious data sets are being used to train artificial-intelligence models that are designed to predict people’s risk of stroke and diabetes, researchers report in a preprint1 on medRxiv. Some of the models seem to have been used in clinical settings, although it’s not clear whether this has led to flawed diagnoses. […]

    https://rbfirehose.com/2026/04/22/nature-dozens-of-ai-disease-prediction-models-were-trained-on-dubious-data/
  5. Nature: Dozens of AI disease-prediction models were trained on dubious data. “Dubious data sets are being used to train artificial-intelligence models that are designed to predict people’s risk of stroke and diabetes, researchers report in a preprint1 on medRxiv. Some of the models seem to have been used in clinical settings, although it’s not clear whether this has led to flawed diagnoses. […]

    https://rbfirehose.com/2026/04/22/nature-dozens-of-ai-disease-prediction-models-were-trained-on-dubious-data/
  6. A very disturbing, and very credible report.

    AI got the blame for the Iran school bombing. The truth is far more worrying
    theguardian.com/news/2026/mar/

    #war #automation #BadData can lead to #WarCrimes - such as intentionally striking a school - but "no one" is responsible?

  7. A very disturbing, and very credible report.

    AI got the blame for the Iran school bombing. The truth is far more worrying
    theguardian.com/news/2026/mar/

    #war #automation #BadData can lead to #WarCrimes - such as intentionally striking a school - but "no one" is responsible?

  8. A very disturbing, and very credible report.

    AI got the blame for the Iran school bombing. The truth is far more worrying
    theguardian.com/news/2026/mar/

    #war #automation #BadData can lead to #WarCrimes - such as intentionally striking a school - but "no one" is responsible?

  9. A very disturbing, and very credible report.

    AI got the blame for the Iran school bombing. The truth is far more worrying
    theguardian.com/news/2026/mar/

    #war #automation #BadData can lead to #WarCrimes - such as intentionally striking a school - but "no one" is responsible?

  10. A very disturbing, and very credible report.

    AI got the blame for the Iran school bombing. The truth is far more worrying
    theguardian.com/news/2026/mar/

    #war #automation #BadData can lead to #WarCrimes - such as intentionally striking a school - but "no one" is responsible?

  11. are trying to advertise some streaming sport event to me, Hornets vs Spurs. In what mad world are (a team) playing (an team)? That's an impressive cock-up.

  12. #Amazon are trying to advertise some streaming sport event to me, Hornets vs Spurs. In what mad world are #RochdaleHornets (a #RugbyLeague team) playing #TottenhamHotspur (an #AssociationFootball team)? That's an impressive cock-up.

    #BadData #Incompetence

  13. #Amazon are trying to advertise some streaming sport event to me, Hornets vs Spurs. In what mad world are #RochdaleHornets (a #RugbyLeague team) playing #TottenhamHotspur (an #AssociationFootball team)? That's an impressive cock-up.

    #BadData #Incompetence

  14. #Amazon are trying to advertise some streaming sport event to me, Hornets vs Spurs. In what mad world are #RochdaleHornets (a #RugbyLeague team) playing #TottenhamHotspur (an #AssociationFootball team)? That's an impressive cock-up.

    #BadData #Incompetence

  15. #Amazon are trying to advertise some streaming sport event to me, Hornets vs Spurs. In what mad world are #RochdaleHornets (a #RugbyLeague team) playing #TottenhamHotspur (an #AssociationFootball team)? That's an impressive cock-up.

    #BadData #Incompetence

  16. It's interesting. After every Teams meeting a popup appears asking 'how was the quality of the audio/video?' along with a five star rating. I always give it five stars, not because it was great, but because if you give it anything other than five stars it grills you to have you say exactly what was wrong with it. #darkpatterns #baddata

  17. It's interesting. After every Teams meeting a popup appears asking 'how was the quality of the audio/video?' along with a five star rating. I always give it five stars, not because it was great, but because if you give it anything other than five stars it grills you to have you say exactly what was wrong with it. #darkpatterns #baddata

  18. It's interesting. After every Teams meeting a popup appears asking 'how was the quality of the audio/video?' along with a five star rating. I always give it five stars, not because it was great, but because if you give it anything other than five stars it grills you to have you say exactly what was wrong with it. #darkpatterns #baddata

  19. It's interesting. After every Teams meeting a popup appears asking 'how was the quality of the audio/video?' along with a five star rating. I always give it five stars, not because it was great, but because if you give it anything other than five stars it grills you to have you say exactly what was wrong with it. #darkpatterns #baddata

  20. It's interesting. After every Teams meeting a popup appears asking 'how was the quality of the audio/video?' along with a five star rating. I always give it five stars, not because it was great, but because if you give it anything other than five stars it grills you to have you say exactly what was wrong with it. #darkpatterns #baddata

  21. I used to be in a team that was responsible for the education and training questions, then in a team that published labour market data, then moved onto compendia publications.
    Both the LFS and the Census were key data for much of my career.
    I hadn't kept up with the post covid changes - and am very sad to see how it has gone downhill.
    But not surprised it failed during a government into cost cutting but not transparency.

    #BadData
    equals
    #Bad Government

  22. I used to be in a team that was responsible for the education and training questions, then in a team that published labour market data, then moved onto compendia publications.
    Both the LFS and the Census were key data for much of my career.
    I hadn't kept up with the post covid changes - and am very sad to see how it has gone downhill.
    But not surprised it failed during a government into cost cutting but not transparency.

    #BadData
    equals
    #Bad Government

  23. I used to be in a team that was responsible for the education and training questions, then in a team that published labour market data, then moved onto compendia publications.
    Both the LFS and the Census were key data for much of my career.
    I hadn't kept up with the post covid changes - and am very sad to see how it has gone downhill.
    But not surprised it failed during a government into cost cutting but not transparency.

    #BadData
    equals
    #Bad Government

  24. I used to be in a team that was responsible for the education and training questions, then in a team that published labour market data, then moved onto compendia publications.
    Both the LFS and the Census were key data for much of my career.
    I hadn't kept up with the post covid changes - and am very sad to see how it has gone downhill.
    But not surprised it failed during a government into cost cutting but not transparency.

    #BadData
    equals
    #Bad Government

  25. #BadData #Bad Analysis
    "In most cases differences between dietary groups were not statistically significant. However, small numbers of vegan cats affected by disorders may have prevented the detection of statistically significant differences between diet groups..."
    Yet:
    "Jointly considering seven general indicators of health and 22 specific health disorders, cats fed vegan diets tended to be healthier than those fed meat-based diets."

    Vegan versus meat-based cat food
    journals.plos.org/plosone/arti

  26. Donc finalement c'est BEAUCOUP de mauvaises données, il manque des profs un peu mais c'est une question dans les centaines, pas les milliers. À se demander à quel point ça arrive souvent que le gouvernement a aucune idée de ce qui se passe sur le terrain

    #Quebec #BadData #Education

    lapresse.ca/actualites/educati

  27. Donc finalement c'est BEAUCOUP de mauvaises données, il manque des profs un peu mais c'est une question dans les centaines, pas les milliers. À se demander à quel point ça arrive souvent que le gouvernement a aucune idée de ce qui se passe sur le terrain

    #Quebec #BadData #Education

    lapresse.ca/actualites/educati

  28. Donc finalement c'est BEAUCOUP de mauvaises données, il manque des profs un peu mais c'est une question dans les centaines, pas les milliers. À se demander à quel point ça arrive souvent que le gouvernement a aucune idée de ce qui se passe sur le terrain

    #Quebec #BadData #Education

    lapresse.ca/actualites/educati

  29. Donc finalement c'est BEAUCOUP de mauvaises données, il manque des profs un peu mais c'est une question dans les centaines, pas les milliers. À se demander à quel point ça arrive souvent que le gouvernement a aucune idée de ce qui se passe sur le terrain

    #Quebec #BadData #Education

    lapresse.ca/actualites/educati

  30. Donc finalement c'est BEAUCOUP de mauvaises données, il manque des profs un peu mais c'est une question dans les centaines, pas les milliers. À se demander à quel point ça arrive souvent que le gouvernement a aucune idée de ce qui se passe sur le terrain

    #Quebec #BadData #Education

    lapresse.ca/actualites/educati

  31. Whomever entered these songs’ ID3 Tags is an asshole:

    (Image: I list of songs in a digital compilation with the Title field showing the artist and title, not just the title, and the Artist Field showing “Various Artists”)

    Also, tell #Bandcamp to stop letting artists write their artist field in all caps: It fucks with automated library filing trying to keep one artist in one folder.

    #music #badData #Assholes

  32. Whomever entered these songs’ ID3 Tags is an asshole:

    (Image: I list of songs in a digital compilation with the Title field showing the artist and title, not just the title, and the Artist Field showing “Various Artists”)

    Also, tell #Bandcamp to stop letting artists write their artist field in all caps: It fucks with automated library filing trying to keep one artist in one folder.

    #music #badData #Assholes

  33. Whomever entered these songs’ ID3 Tags is an asshole:

    (Image: I list of songs in a digital compilation with the Title field showing the artist and title, not just the title, and the Artist Field showing “Various Artists”)

    Also, tell #Bandcamp to stop letting artists write their artist field in all caps: It fucks with automated library filing trying to keep one artist in one folder.

    #music #badData #Assholes

  34. Whomever entered these songs’ ID3 Tags is an asshole:

    (Image: I list of songs in a digital compilation with the Title field showing the artist and title, not just the title, and the Artist Field showing “Various Artists”)

    Also, tell to stop letting artists write their artist field in all caps: It fucks with automated library filing trying to keep one artist in one folder.

  35. BAD DATA
    An informative and apolitical new book reminds us that statistics are not always what they seem.

    How Do They Know This?

    Important insights about #statistics that the public, all scientists and all politicians should know about !!

    quillette.com/2022/12/08/how-d

    #sociology #sociologist @sociology #politicalscience #econometricians #publicstatistics #PolicyResearch #psychology #policymaking #data #bigdata #baddata @politicalscience #indicators #policyindicators #socialindicators

  36. BAD DATA
    An informative and apolitical new book reminds us that statistics are not always what they seem.

    How Do They Know This?

    Important insights about #statistics that the public, all scientists and all politicians should know about !!

    quillette.com/2022/12/08/how-d

    #sociology #sociologist @sociology #politicalscience #econometricians #publicstatistics #PolicyResearch #psychology #policymaking #data #bigdata #baddata @politicalscience #indicators #policyindicators #socialindicators

  37. BAD DATA
    An informative and apolitical new book reminds us that statistics are not always what they seem.

    How Do They Know This?

    Important insights about #statistics that the public, all scientists and all politicians should know about !!

    quillette.com/2022/12/08/how-d

    #sociology #sociologist @sociology #politicalscience #econometricians #publicstatistics #PolicyResearch #psychology #policymaking #data #bigdata #baddata @politicalscience #indicators #policyindicators #socialindicators

  38. BAD DATA
    An informative and apolitical new book reminds us that statistics are not always what they seem.

    How Do They Know This?

    Important insights about #statistics that the public, all scientists and all politicians should know about !!

    quillette.com/2022/12/08/how-d

    #sociology #sociologist @sociology #politicalscience #econometricians #publicstatistics #PolicyResearch #psychology #policymaking #data #bigdata #baddata @politicalscience #indicators #policyindicators #socialindicators

  39. BAD DATA
    An informative and apolitical new book reminds us that statistics are not always what they seem.

    How Do They Know This?

    Important insights about #statistics that the public, all scientists and all politicians should know about !!

    quillette.com/2022/12/08/how-d

    #sociology #sociologist @sociology #politicalscience #econometricians #publicstatistics #PolicyResearch #psychology #policymaking #data #bigdata #baddata @politicalscience #indicators #policyindicators #socialindicators

  40. So instead of everyone being distracted by an AI that can mimic biblical verse/language for mundane or amusing questions - perhaps we should start asking the questions which matter - how can we stop AIs being built and trained on bad data?

    #gdpr #compliance #ethics #bias #law #humanimpact #responsibletech #risk #security #humanrights #machinelearning #deeplearning #data #baddata