Is this WhatsApp AI antitrust probe missing the core privacy threat?
I’m not an AI specialist nor a WhatsApp user, but antitrust looks to me like the least of the problems here.
“Meta to face competition probe over AI use in WhatsApp, Ribera confirms.” —Euractiv
If there’s an AI feature working inside WhatsApp, isn’t it reasonable to assume that at least parts of the conversation are leaving the app? After all, it has to be analyzed by a model sitting in a server somewhere, undermining the end-to-end encryption Meta is constantly bragging about?
And whatever Meta is claiming to do to keep the conversations encrypted will never convince me. If history can tell us something about this company, it is how irresponsible they are with users' data.
iPhone market share is much smaller than I thought
I found the Hey there! You are using WhatsApp: Enumerating Three Billion Accounts for Security and Privacy a fascinating read. Unfortunately, it confirmed once more my perception of how Meta couldn’t care less about users' personal information. How is it possible that a single server could download all that data? Unbelievable!
Because of my love for numbers and graphs, I also spent a lot of time on the ones in the paper that, by the way, go way beyond WhatsApp. For example, the one below has to be the best representation so far of the Android-iPhone market share. I’m not a statistician, but I believe a 3.5-billion sample can generate a pretty accurate representation of reality.
To be completely honest, I thought Apple had a bigger market share in both the US and Europe, and it would be cool to see how things changed in a year or two. Meta said they patched the data downloading issue, but given their history of data leaking, I’d say there’s a good chance we’ll be able to see an updated graph in the near future.
3,5 Bilhões de Contas—entenda a falha de segurança do WhatsApp
Um grupo de pesquisadores conseguiu descobrir 3,5 bilhões de usuários do WhatsApp e baixar a imagem de perfil, bio, data de criação e atualização das contas, em que telefones e computadores as contas são usadas, e todas as chaves públicas de criptografia. Com estes e outros dados, conseguiram descobrir diversas outras informações dos usuários.
Tenho visto vários comentários dizendo que o vazamento não foi tão grave assim porque o conteúdo já era público. Minha opinião é um pouco diferente. No episódio de hoje, explico o estudo e o que penso a respeito desta coleta massiva de dados que foi completamente ignorada pela Meta.
Tudo que foi feito está publicado num Paper repleto de gráficos e análises interessantíssimas. Por exemplo, o gráfico abaixo demonstra o share de mercado do Android e iOS em todas as regiões do mundo.