Mark Harrison

Dec 27, 2024 | acrjournals.onlinelibrary.wiley.com | Megan Thomas |Mark Harrison |Cheryl Barnabe |Charlene Ronquillo |J. Antonio Avina‐Zubieta |Anna Samson | +2 more

Supporting Information Filename Description acr25487-sup-0001-Disclosureform.pdfPDF document, 438 KB Disclosure Form acr25487-sup-0002-Supinfo.docxWord 2007 document , 25.8 KB Appendix S1: Supporting Information

Jun 3, 2024 | medium.com | Mark Harrison

Articles about AI risks tend to be: sensationalist tropes (the robots from the future are coming to kill us - follow me if you want to live) hype-busters (ooh, LLMs hallucinate as much as humans do) worries about bias (ooh, LLMs share the prejudices of the data they were trained on as much as humans do) However, we shouldn't get carried away by these risks to the extent we treat ML deployments as "a different category" from any other deployment.