Подтвердите e-mail

Для публикаций, комментариев, реакций и сообщений подтвердите адрес.

←

Публикация

Предыдущая публикация
Institute for Replication

Thanks for the feedback and checking our paper. We are happy to issue a correction if coauthors did not mention competing interest. Many drafts of the papers circulated to all coauthors and we asked them to release COI if necessary. Only a few did as you note...

110
Institute for Replication

Concerning the main coauthors at I4R (those who mostly did the empirical analysis and conception), we have no COI related to AI groups. We have no existing or previous relationship with AI/tech groups. Our funding for this project and other projects come almost entirely from Coefficient Giving...

Обсуждение

1 прямых ответов · 6 сообщений

Institute for Replication

SSHRC and previously Sloan. We wrote "This research and AI replication games were funded by Coefficient Giving project “Benchmarking LLM agents on real-world tasks: Reproducibility” and the Alfred P. Sloan Foundation Foundation grant G-2023-22326. We also benefited from funding to host games from...

Ответ для Institute for Replication
Institute for Replication

the Universities of Toronto, Ottawa, Cornell, and Tilburg." On a personal note, views about AI at I4R vary widely from Abel (chair) disliking AI to Bruno (researcher) loving it. We will email once more all authors about the possibility to release a corrected COS, cc'ing you tomorrow. ...

Ответ для Institute for Replication
Institute for Replication

Also, none of us (core group in Canada) is getting paid by anyone other than uOttawa. No $$ for doing replication other than our base salary at uOttawa. My (Abel) salary is publicly available on sunshine list: www.canadasunshinelist.ca/ontario/empl.... I made a bit more than that ...

Ответ для Institute for Replication
Institute for Replication

through referee reports and other services, but that's pretty much it. Happy to jump on a zoom call and show you my tax reports. The other coauthors handling most of the data analysis are my postdoc fellows and PhD student.

Ответ для Institute for Replication
Joe Bak-Coleman

I appreciate the thorough response. It would be worth considering where coefficient giving gets that giant mountain of cash, and the various advocacy they engage in w/r/t AI. In the past, industries have used nonprofits to fund favorable work, laundering in a sense the apparent conflict.

Ответ для Joe Bak-Coleman
Joe Bak-Coleman

Given that history it’s something I’d feel compelled to declare, or at least discuss with editors. Beyond that, I would find it very unlikely that none of the authors have received funding from, have worked at, etc etc… any tech company. Possible but doubt it!