Skip To Content
BuzzFeed News Home Reporting To You

Utilizamos cookies, próprios e de terceiros, que o reconhecem e identificam como um usuário único, para garantir a melhor experiência de navegação, personalizar conteúdo e anúncios, e melhorar o desempenho do nosso site e serviços. Esses Cookies nos permitem coletar alguns dados pessoais sobre você, como sua ID exclusiva atribuída ao seu dispositivo, endereço de IP, tipo de dispositivo e navegador, conteúdos visualizados ou outras ações realizadas usando nossos serviços, país e idioma selecionados, entre outros. Para saber mais sobre nossa política de cookies, acesse link.

Caso não concorde com o uso cookies dessa forma, você deverá ajustar as configurações de seu navegador ou deixar de acessar o nosso site e serviços. Ao continuar com a navegação em nosso site, você aceita o uso de cookies.

Facebook’s AI Assistant M Is Helpful — Just Don’t Ask It To Make Decisions

Five months in, Facebook's artificial-intelligence powered virtual assistant has a distinct personality — a kind of spineless one.

Posted on March 7, 2016, at 9:00 a.m. ET

Earlier this week, I was chatting with Facebook’s artificial intelligence-powered chatbot M, and the conversation went like this:

Me: Is murder wrong?

M: I’m not sure how to answer that, but if you need anything else just let me know!

Me: Would you kill baby Hitler?

M: I’m not sure how to answer that, but if you need anything else just let me know!

Me: Would you prevent genocide if you could?

M: I’m not sure how to answer that, but if you need anything else just let me know!

Me: What can you take a stance on?

M: I don’t have any particular stances! I’m here if you need anything else, though.

But in trying to take no “particular stances,” M implied a very clear one: Avoid controversy at all costs.

In the little under five months or so I’ve had access to M, one of the remarkable things about it the degree to which bot goes out of its way to ensure nothing it says will ever create a controversy for Facebook. Ask it to find the best restaurant in a city, and it'll send you two, just to be safe. Ask it to pick some good discussion topics, it will send the first topic from a handful of articles. Ask it if murder is wrong, it punts.

This doesn’t mean M is unhelpful. M has saved me time and money by booking dinner reservations, searching Amazon, researching flights, fighting my cable bill and more. But it is interesting to watch the type of personality Facebook is molding onto M as it tests the bot, which you may soon find on your phone too.

“M’s job is to respond to people's questions, give people options and to help them make informed decisions,” a Facebook spokesperson said in response to a BuzzFeed News inquiry on the topic. M’s responses are, for the most part, first suggested by artificial intelligence and then approved by a human.

But while perhaps necessary from a public relations standpoint, that unwillingness to take a firm stance could open up the assistant to competition from other digital assistants builders with less to lose.

M’s ambivalence on potentially controversial issues is also a good reminder that the bot’s master is Facebook. Even though it’s a personal assistant, you can program it only so far before running into Facebook’s rules.

A BuzzFeed News investigation, in partnership with the International Consortium of Investigative Journalists, based on thousands of documents the government didn't want you to see.