Google is in the midst of giving Gmail a revamp. The world’s most popular email system hasn’t had a facelift in over a decade and, frankly, was looking rather tired and out-of-date.

There has been much criticism about the redesign. It has annoyed many die-hard fans of Gmail who have become used to the system because it has changed so little over the years. One thing Google has not learned is that if you don’t change things on an on-going basis, users get so wedded to them that they will hate anything you do to improve things. For instance, take a look at Amazon a year ago and you’ll see it is significantly different to the look and feel of today. However, people are not so vociferous in their complaints. And that’s because they haven’t noticed the incremental changes. When you change something suddenly, it grates, and people do not like it. The folks at Google need to take a lesson from Amazon; make small changes every day, rather than massive changes every decade.

On top of the sudden – and frankly retrograde – redesign, Gmail has new functions which Google is very proudly shouting about. These new functions are based on artificial intelligence (AI) and what appears to have eluded the search giant in that is the word “artificial”; in other words, this intelligence is not real.

Here’s a video of some Google staffers waxing lyrical about one of the new AI features “Smart Reply”.

What you notice from this video is that the reply isn’t smart, but dumb. Here’s why. Email is a conversation. In real-world conversations, if you were invited to a party, as shown in the video, you would respond with more than a couple of words. You’d ask questions, like “who else is going” and “where is it” and so on. You’d chat with your friend and you’d get all the information within a minute. With “Smart Reply” Gmail will send a couple of words. To which the recipient’s “Smart Reply” will respond with a single word “Thanks” or something like that. Now you need more information, so you need to generate more emails. Far from saving time, Smart Reply is going to generate more emails and more time wasting. Bonkers. You’ll have a tough job of actually achieving “inbox zero” when you keep getting so many nonsense responses.

Here’s an example. I received an email from a software company in response to a support request. The email asked for a screenshot of the issue and further clarification of the problem. Here are the options that Gmail’s “Smart Reply” gave me.

None of these deal with the question I was asked. If I pressed any of these buttons all it would do is generate more emailing between me and the software company.

Here’s another instance. I am due to speak at a conference in a couple of months and the organisers have set up a website with information about me and my talk. The email sent me a link and asked me to check the page and let them know if I wanted anything changing. Here’s what Gmail’s “Smart Reply” suggested I send.

Again, none of these options deals with the questions asked in the message I was sent.

Google has clearly realised its AI system is far from perfect. It has just announced that it will be providing Gmail users with the option of switching it off. What this means, though, is that Google is making email more complex, rather than simpler, because of an array of settings and options you need to configure. Just imagine if you had to do all that with real-world conversations…!

Gmail’s artificial nonsense goes further

It’s not just “Smart Replies” that Google is introducing with its new Gmail. It is also using AI to provide something it calls “Smart Compose“. This is where the system actually starts writing parts of your email for you.

But I have a question for you. When you are holding a conversation, how much do you like it when the people you are speaking to finish your sentences? Like Google’s new system, they too are using a neural network to predict what you want to say and saying it before you get a chance. Smart Compose is going to be just as annoying as your colleagues and friends who try to finish your sentences.

Furthermore, it is going to lead to massive increases in poor communication. In the rush of the day-to-day office turmoil, many people will accept the suggestions because they are pressured. Those suggestions will not always be valuable, useful or correct. This will result in incorrect actions taking place and increased amounts of email to correct things or to clarify things because of confusion generated by the so-called “smart” system.

I hope Google is saving money for the inevitable lawsuits when things go wrong because someone acted on an email which was generated using Smart Compose. The “smart” part of the message might cost the recipients significant harm if the AI did not fully appreciate what was going on and offered a suggestion which was not appropriate.

AI is not the answer

Google, like many tech companies, is so in love with technology and its possibilities they live in a “bubble” which is far from the real world. This became apparent to me when I had to visit a Google office in Central London. The address I had been given was clear and I knew roughly where it was as I used to work not far from that part of London several years ago. I arrived in the square but could not see the building. There were a number of small tower-blocks. One was from Barclays Bank, another was a law firm. There were some cafés and smaller offices too. But I could not see either the Google logo or the number of the building. Eventually, by a process of elimination, I realised which building it was. I went in and went up to the meeting room where I asked if the lack of signage outside was some kind of security issue. The staffers looked at me in a rather puzzled fashion and said: “but don’t you just use your phone to find the building?” Well, no, I don’t. I look around me. I use my eyes.

What this little incident showed me was that these people are so wedded to their technology they don’t appreciate the rest of the world does not live in their little bubble. Yes, I do use my phone. Yes, I do use Google Maps. Yes, I am also a technology fan. But I am also human…!

After all, what happens if my mobile phone is broken, or the connection drops, or the sun is shining so brightly I cannot see the screen? If I relied on technology and its artificial intelligence to such a degree I’d be lost.

There is no doubt that AI can help. Already Gmail uses some AI to sort out all those promotional emails and separate them from the real emails you want to see. But sometimes you can take AI too far. Just because something can be engineered does not mean it needs to be done. We can engineer it, for instance, to eliminate men because we can continue the human race with artificial insemination. Doesn’t mean it’s a good idea.

Similarly, we can engineer it for computers to write our emails. It doesn’t mean it’s a good idea either.