Morgan Gregory

@morganjgregory

Google CTO Office. Adventures in Responsible AI and raising 2 girls under 5. Opinions are my own.

Vrijeme pridruživanja: rujan 2009.

Tweetovi

Blokirali ste korisnika/cu @morganjgregory

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @morganjgregory

  1. Prikvačeni tweet
    22. sij

    Heading to the today to talk about I can't share slides, but I've posted links to resources and papers here:

    Poništi
  2. 28. sij

    We need more diversity in tech, initiatives like this (that also recognize the need for diversity along other dimensions) help move us in the right direction

    Prikaži ovu nit
    Poništi
  3. 28. sij

    This is great: "...will focus on increasing the percentages of women who graduate with computing degrees and who get hired into (and retained/promoted at) tech jobs; it also aims to increase the amount of venture capital invested in startups with at least one female founder."

    Prikaži ovu nit
    Poništi
  4. 28. sij

    Dr Ayanna Howard shared fascinating findings on how people build trust with robots, how robot errors impact trust building, and how the time at which robots make errors impact trust over time

    Poništi
  5. 16. sij

    Cont: I am so inspired by all the great work these people are doing on fairness in AI, and this is just to name a few.

    Prikaži ovu nit
    Poništi
  6. 16. sij

    I'm excited to share an article I wrote on fairness in AI. It's a nuanced and complex topic and I'm hoping to provide an accessible starting point. Other great voices on this topic:

    Prikaži ovu nit
    Poništi
  7. 14. sij

    I can relate! This is me trying to pick a flavor at an ice cream shop

    Poništi
  8. 10. sij

    Constantly inspired by the incredible work of the people around me. On Ethical AI: "Our goals are to apply the best currently-known techniques in these areas to our work, and also to do research to continue to advance the state of the art"

    Poništi
  9. 7. sij
    Prikaži ovu nit
    Poništi
  10. 7. sij

    When a biomedical researcher broke agreed to ethics rules for using CRISPR, he was banned from conferences, scientific journals, and sent to jail for 3 years. to students at the seminar - should we have similar rules and consequences in AI?

    Prikaži ovu nit
    Poništi
  11. 7. sij

    "When someone says let's do good, no one calls on the Computer Scientists When someone says let's do AI for Good, it's ONLY Computer Scientists" at the Seminar last night on the importance of multi-disciplinary teams in AI

    Poništi
  12. 5. sij

    The thread kicked off an investigation by financial regulators, so answers should be coming (fingers crossed).

    Prikaži ovu nit
    Poništi
  13. 5. sij

    To be clear, no one has published a study on the algorithm used by either Apple or Goldman Sachs to determine credit worthiness or credit limits, so the accusations that it discriminates against women can’t be proved or disproved at this time.

    Prikaži ovu nit
    Poništi
  14. 5. sij

    It’s not surprising this thread got the attention it did. What this points to is that people are increasingly concerned about fairness in AI, and are increasingly aware that historical societal bias can be baked into an algorithm through the data we train that algorithm with.

    Prikaži ovu nit
    Poništi
  15. 5. sij

    To me, the most interesting thing about this thread was the strong desire from many for transparency into the credit algorithm, and the inability for Apple or Goldman Sachs to provide it.

    Prikaži ovu nit
    Poništi
  16. 29. pro 2019.

    In 2019 I discovered my passion for when invited me to the Thank you for the crash course intro: - bias in healthcare, - bias in AI, Cynthia Dwork - differential privacy

    Poništi
  17. 17. stu 2019.

    One of those things I didn't realize was Canadian before I moved to the states. Also in that camp: Smarties, Kraft dinner, and the words garburator and washroom.

    Poništi
  18. 24. lis 2019.

    By using projected treatment costs as a proxy for how sick a patient was, this algorithm discriminated against black patients. What a powerful reminder that data sets used for training AI can have harmful hidden bias baked in

    Poništi
  19. 21. lis 2019.

    I was honored to be able to share the amazing work of 4 teams at the summit in Seattle. It's so inspiring to see what they achieved by applying AI to scientific discovery, and co-innovating on Google Cloud.

    Poništi
  20. 12. lis 2019.

    "2.2% of the 130B of venture dollars last year went to female founders" thank you for not shying away from this topic - the more we can do to raise awareness, the better

    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·