Experimental browser for the Atmosphere
So what? 1. Multilingual evaluation must be standard, not optional. 2. Domain-aware fine-tuning & transparent training data are overdue. 3. Regulators and platforms should consider demanding audits before health-chatbots go live. Full paper at arxiv.org/abs/2504.18310
Apr 28, 2025, 10:06 AM
{ "uri": "at://did:plc:aj3d62z2hv6fh5eh5cw4zia6/app.bsky.feed.post/3lnujew5els2o", "cid": "bafyreicofljwocoeeqbv4zoixfnpdhtnuh4uolyrytzr6fq6id7ntpsvpu", "value": { "text": "So what?\n\n1. Multilingual evaluation must be standard, not optional.\n2. Domain-aware fine-tuning & transparent training data are overdue.\n3. Regulators and platforms should consider demanding audits before health-chatbots go live.\n\nFull paper at arxiv.org/abs/2504.18310", "$type": "app.bsky.feed.post", "embed": { "$type": "app.bsky.embed.external", "external": { "uri": "https://arxiv.org/abs/2504.18310", "thumb": { "$type": "blob", "ref": { "$link": "bafkreihk7eontgs6o5bbufzem4xbbpzcam3lsixrwinngx243ilmregbr4" }, "mimeType": "image/jpeg", "size": 147062 }, "title": "Artificial Intelligence health advice accuracy varies across languages and contexts", "description": "Using basic health statements authorized by UK and EU registers and 9,100 journalist-vetted public-health assertions on topics such as abortion, COVID-19 and politics from sources ranging from peer-re..." } }, "langs": [ "en" ], "reply": { "root": { "cid": "bafyreifff54y6i4uu7svol3qxxoxmjlllscfavnaolhfb4emuvqbzamvfi", "uri": "at://did:plc:aj3d62z2hv6fh5eh5cw4zia6/app.bsky.feed.post/3lnujeiqd422o" }, "parent": { "cid": "bafyreib2b3iyks5cygxvja4zblazxnnngjb5e2ozmodv57zkvovfpwg5pe", "uri": "at://did:plc:aj3d62z2hv6fh5eh5cw4zia6/app.bsky.feed.post/3lnujesrock2o" } }, "facets": [ { "index": { "byteEnd": 270, "byteStart": 246 }, "features": [ { "uri": "https://arxiv.org/abs/2504.18310", "$type": "app.bsky.richtext.facet#link" } ] } ], "createdAt": "2025-04-28T10:06:03.563Z" } }