- With Smart Reply, Gmail suggests three answers to every email we receive.
- The feature uses machine learning, refining its proposals as it is used.
- Will artificial intelligence end up completely replacing us in digital exchanges?
“Well received”, “It is noted”, “Thank you very much”. It is laconic but effective and rather relevant.
For several months, users of Gmail have the strange sensation of being watched by a small curious who enjoys reading any mail from his virtual couch.
Helped by artificial intelligence, Google’s messaging has, very gradually, made us chew the job.
She offers three ready-made answers to each of our emails, she finishes the end of our sentences in full writing and she reminds us to answer our messages. No need to think, the machine does it for us.
“Smart reply uses machine learning to provide more and more relevant answers as you use it. If you are more “thank you!” Than “thank you”, we will propose answers that are more like you “, explained Google in 2017.
Halfway between an assistant and a digital double, the algorithm flies downright our good behavior, in reminding us when we forgot to answer someone.
Practical but a bit scary. What do these features predict for the future of our email exchanges (considering that digital missives will be part of our future)?
Then How About Our Personal Data?
Technology seems to know the content of our private exchanges. Worse, it does not even seem to hide it.It is to ask what is the true meaning of the term “personal data”.
“From the moment we entrust our messages to a private operator, it can certainly be accessed, but we can not imagine that a person is behind because it would be much too expensive,” said Jean-Jacques, amused.
Gabriel Ganascia, researcher in artificial intelligence, president of the CNRS ethics committee. It is not an intrusion into individual life in the sense that it is treated en masse, “he contradicts. And Google confirms. “We do not read e-mails from our users or from our customers.”
If, for the time being, the algorithm of the giant of the Web does not look very far (the answers are rather crude: heard, thank you very much, it is noted, see you soon), it gives a foretaste of what is waiting for us in the era of AI.
Goodbye taking the administrative head, hello intelligent organization of emails and automated writing (the proof with “smart compound” launched last October by the firm of Mountain View)
Nothing very spectacular, you say? “Regardless of the technical difficulty, we must avoid giving the feeling of an intrusion into the autonomy of the user,” says Jean-Gabriel Ganascia.
You must not offend your susceptibility nor fan your paranoia. We remember the madness that surrounded the release of Google Glass.
In 2013, the Stop the Cyborg campaign encouraged people to “call for free, unguarded areas so they can talk and behave freely.” Google’s glasses were perceived as Big Brother in power.
You Have To See Artificial Intelligence As A Partner
With Smart Reply, Gmail flirts with the limit of the acceptable. Fortunately, messaging only suggests, and the most annoyed can disable the function in the settings.
“I can not imagine that one day, there is a substitution, the experience shows that we must see artificial intelligence as a partner,” insists the specialist.
We will write less, but we may spend more time on important things. “Artificial intelligence will have a big effect on professional letters, they will be pre-written, but private correspondence, I do not think.
It’s a bit the same as with conversational agents, it can work as long as the semantics are unambiguous, unambiguous, and play on words, “continues Jean-Gabriel Ganascia.
But in everyday life, we play on the double meaning and it is very difficult to imitate for a machine.
Today, the trend is to video. The future of the exchanges could be played on this ground and less on the mails per se.
We would film our lives and the artificial intelligence would be in charge of selecting the most striking events. “You saw something incredible in the subway and you ask the machine to review this sequence”, imagine the specialist.
We would send it to our loved ones or post it on social networks very easily. It is playable, but the artificial intelligence is not yet robust enough and, above all, it will be necessary to accept to have a camera screwed on the head permanently to film our daily life.
Who knows, the Google Glass may not be completely dead … But for now, we will stick to: “OK, thank you! “