Have you sent a message recently using a smart phone? I have. I’m sure you have too. When you use any social network or app chances are that one of the big Silicon Valley giants already thinks they know what you want to send. It’s incredible how the many tools like Gmail, LinkedIn, Facebook and Instagram become much more than just tools but prescriptive automation machines.
Google and LinkedIn recently updated their messengers to become more suggestive. You see smart replies everywhere. Here are just a few examples.
A few months ago Google took this to the next level by releasing Smart Compose. The Silicon Valley giant now predicts the body of your email step by step. Google says “Compose, a new feature powered by artificial intelligence, to help you draft emails from scratch, faster.”
One one hand it’s nice that Google tries to help me compose my message on the other hand that suggestion-based composing makes me feel less in charge of what I’m writing and consequentially what I’m thinking. Maybe I didn’t want to ask Jerry to meet up on Thursday. The first release of smart compose is still in it’s infancy but at the speed at which Google is able to evolve its products I’m sure it’ll be only a few months before the suggestions in Smart Compose become more pervasive and nuanced. At some points I won’t be sure if I decide what to type into the machine or if it’s the other way around. So, the question is how far can AI algorithms go in replacing the human touch of a message and what are the consequences?
How Far Will AI Smart-Replies Go?
No doubt, being able to respond to a message by the click of a button is a great time-saver and if the advanced algorithms of Google and co. learn from my adjustments that I make to their responses to make them sound even more like me, even better. I was curious to see if someone could most of their daily routine with Smart Replies? Well, here’s Mike Brown from Inverse.com who had a go at it. Visit his post to find out how is smart-reply-only policy worked and how his friends and colleagues reacted.
For me that’s a nice experiment but not the main point. It’s more important to understand how much influence these smart replies have on our freedom to write our own messages as compared to what was suggested. After all, the currency for big tech is attention. Google and others in Silicon Valley make their money by getting eyeballs on ads. Their machine learning and AI efforts are heavily steered to make that happen. My question is, is this innocent feature a wolf in a sheepskin?
But am I just and exaggerating tech-pessimist? Let’s try and think this through. Whilst, I’m hugely enthusiastic about what great innovation can do for humanity, I would always try to see both sides and understand the underlying motives of the companies that give us stuff for free. How about if LinkedIn, Facebook and Google could direct our attention to an advert on the web using smart replies. I’ve shown you this picture before…
What if Google suggested to meet at a specific coffee place? That coffee place might have paid for a smart compose ad to be placed for me. Me being the persona in their algorithms that lives in East London and likes trying new hipster-like coffee places. How long will it be before LinkedIn, Google and Facebook start suggesting the restaurants, cafes or shops where I want to meet Jerry and other people?
It might look something like this.
I’m sure it’s a turnkey solution for Google which already has my location, knows where I live, where I drink my coffee and eat my dinner. And that point my smart phone is that smart that it takes over not just the composing of emails but takes even decisions for me. Will I still be free to choose? Or will the advertising-based economy that runs on me consuming what is suggested take over that part for me? In a very exaggerated dystopian way it reminds me of Black Mirror‘s 15 million merits episode in season 1. Well, that escalated quickly…
But the point still stands, this is very imaginable. And if it can be imagined, it can be done, if it can be done it will be done.
Where does this leave us?
Where does this leave us today? Smart replies are great to save time. But that’s pretty much where it stops. It’s a small step to make Gmail and other messenger apps ad-based money-making machines by not just completing our sentences but by suggesting specific shops, cafes, restaurants to visit and buy from. That will impact our freedom and I’m not sure if I like it.
Hence why, I deactivated the smart compose feature of Gmail. For now at least.
When it comes to AI I’m very much an optimist to what it can do to help humanity. In this specific case, I’m a bit cautious. Giving this much power over what I’m thinking and writing to tech companies is not something I feel comfortable about. Recent news doesn’t help much in eliciting trust in companies like Facebook. It was revealed that Zuckerberg wrote in an internal memo: “That may be good for the world, but it’s not good for us” referring to the need for user data to keep people on Facebook as the Guardian reported.
I guess being open, transparent and building trust is something that’s still lacking in some big tech companies. But it is undoubtedly important for Google, Facebook and others to learn what Spider-Man learnt many years ago already. It’s that with great power comes great responsibility.
Let me know what you think will happen with smart replies in the future in the comments section.