LaMDA, Google and when playing with an artificial intelligence is possible

By now everyone should know: the engineer Blake Lemoine has been suspended from employment and salary at Google for leaking confidential information from the company that supposedly involved the revelation of a singular fact: an artificial intelligence with which the company worked. company would have, according to Lemoine, become self-aware.

A long conversation between the engineer and LaMDA (as AI is called) can be read on thousands of sites on the Net. Futurists are divided between those who believe that it is a matter of time or that the singularity has already been reached (the moment when a machine becomes more intelligent than a person) and those who maintain (given that we are not even remotely clear about what being conscious is or where consciousness is generated) that it is directly impossible for that to happen.

As the days have gone by, the most widespread opinion, and shared by the author of this column, is that it is rather a competent phrase aggregator and not so much a real conscience.

The reason is somewhat pedestrian but logical: if we ask a person if he would murder his mother, the normal thing is not that he answers “yes” or “no”, but “what the hell are you saying?”;

but LaMDA's answers are always complex and flowery variations of “yes” and “no”: that is, there is no room for surprise, improvisation, rejection, those little things that make us people.

This makes us suspect that more than a consciousness that

the questions, it is rather a text generator that formulates

answers .

The fact is that, regardless of whether or not Google has achieved the vaunted singularity with this artificial intelligence, LaMDA's conversation (and the recreations that dozens of youtubers have uploaded to the internet) are very much to the point with an extraordinary video game that appeared on last month.

This is

, a

and sequel to

, which in 2013 shook the pillars of the interactive industry by proposing one of the first metalludic experiences in the video game world.

More information

Don't say 'streamer', say 'joueur en direct': the battle for the language of the future begins

In the game we are an office worker (Stanley) who, in the first person, moves through his company, which has suddenly become empty.

In the manner of Unamuno's

, we start the video game accompanied by a (velvety, surgical, serious) narrator's voice that recounts how we left our cubicle and passed by the tables of our colleagues.

That voice accompanies us in our first steps until we reach a room where there are two open doors... and the narrator says that we choose to cross the one on the left to go up to the director's office.

But the voice narrates our decision

we make it.

From there, freedom.

Our actions can follow the indications of the Voice (performed in an extraordinary way by the actor Kevan Brighting), contradict them, reformulate them... and our behavior will activate complicit, hostile, servile or manipulative reactions of a voice with which we will gain complicity and nuances.

In the case of LaMDA, it is the Google engineer's questions that trigger the responses, but in the case of

it is our own actions that provoke the different interactions of the narrator until triggering one of the twenty-odd possible endings.

Soliloquies.

Direct questions to Stanley or the player himself.

Deep reflections on what it means to exist or play.

It is a voice so well represented, so well interpreted, that after several hours playing the game we may fall into the temptation of thinking that there is something there, that there is something more behind that presence that accompanies us incessantly... until we we realize that no, that everything is a game that follows the guidelines of an extraordinary, yes, creative mind (very human) that is behind the script.

The very interface of 'The Stanley Parable' dialogues with us and sometimes leads us to uncomfortable situations.

Perhaps we are wrong and LaMDA is, in effect and as she declares herself, “a person”.

Or maybe not.

But the important thing is that it is a system so well done that it has been able to fool the Google engineer in charge.

Artificial intelligences (conscious or not) will be increasingly important pieces in the world, and will be behind many things: from educational strategies to phone scams.

It is not a bad idea to learn to recognize them and begin to know how to interact with them.

And in this case, as in many others, education can begin through the interactive medium.

After all, how can the reader know that whoever wrote this is a real person?

Want to read another story from this paper, Stanley?

You can follow BABELIA on