Eliza, Terminator and the Undercover Danger by ChatGPT

It was 1966 when Joseph Weizenbaum, a pioneer of artificial intelligence at MIT, discovered that he had something uncomfortable on his hands.

He had developed one of the first chatbots, a computer capable of faking a human conversation with reasonable success.

Her name was Eliza, and she would become a milestone for that fledgling technology.

But also at a turning point for his inventor, after observing the effect he had on people.

Weizenbaum was horrified after letting his secretary use Eliza: after a while, he asked her to leave the room to have privacy in his conversation with the machine.

The anecdote serves to read with perspective the current phenomenon of ChatGPT.

Eliza, in its Doctor version, had been created as a parody of the responses of psychotherapists who returned the user's statements in the form of questions.

The intent was “to show that communication between a person and a machine is superficial,” but seeing his secretary and many others open up fully when conversing with Eliza, Weizenbaum discovered a very different truth.

"I had not realized that even very short exposures to relatively simple programs could induce powerful wishful thinking in normal people," he later wrote.

Weizenbaum's program was very basic, far from the sophistication of the current conversational intelligences that grab headlines and fill the networks with spectacular examples.

But the effect they cause is the same as Eliza, as demonstrated by the Google engineer convinced that LaMDA, another of these machines, had the consciousness of a seven-year-old child.

We continue to project human capabilities onto machines because we humans are also programmed to chat.

As neuroscientist Mariano Sigman explains, what defines us as a species is that we are conversational animals: we define, shape and fulfill ourselves through the words we share with others.

Dialogue is in our DNA and the brain resolves this cognitive dissonance by accepting that this program, even though we know it is a black silicon box,

After Eliza would come Parry, who pretended to be a schizophrenic, and later Alice or Siri, or the most recent and well-known for the Spanish public, such as Irene from Renfe and Bea from Bankia (virtual assistants developed in Spain).

At the worst of the bank's reputational crisis, with the investigation of Rodrigo Rato for his alleged crimes, it was Bea who kept the bank's website upright with her ability to give small talk to those who came in to saturate its systems. and knock down the portal.

People couldn't resist pouring out their outrage on her with

insults at her.

More information

Commercial interests shape the future of artificial intelligence

Weizenbaum didn't understand why people would take Eliza as the first step toward a machine that could simulate human intelligence.

He thought it was a dangerous fantasy and that it was "monstrously wrong" to understand it as anything more than just a program that performed a function.

Weizenbaum abandoned Eliza and became a critic of the idea that machines could be intelligent, because injecting such a frame of mind into society would be "a slow-acting poison."

The words of this pioneer now resonate with objections like the one made by Emily Bender, who insists on repeating that there is nothing magical about ChatGPT, it is just a parrot.

A very sophisticated parrot and with many readings ("stochastic", she qualifies), but a parrot.

This computational linguist is one of the biggest critics of the promoters of these programs that are already flooding everything.

Tools that will be very useful and that will revolutionize many activities, without a doubt, but that lack regulation and transparency.

Bender demands that the companies that promote these chats stop speaking in the first person, as if they were a conscious being: “They must stop making her appear human.

He shouldn't be speaking in the first person: he's not a person, he's a screen."

“They want to create something that looks more magical than it is,

but in reality it is the machine creating the illusion of being human”, denounces Bender.

“If someone is in the business of selling technology, the more magical it seems, the easier it will be to sell it,” he settles.

It's a commercial trick we can't resist.

Like Geppetto, we want the wooden boy to be a boy

.

Freeing Skynet

programs for all their services

that they had been working on for years, but that they had not dared to spread among users until fashion, hype , arrived

from ChatGPT.

For example, Google is going to include it in its business tool, Workspace, and Microsoft in Office.

This has generated an explosion of interest and also interesting (and interested) criticism, such as the open letter signed on Wednesday by a thousand specialists calling for a six-month moratorium on chatbot development.

“AI labs have entered into an uncontrollable race,” they denounce, “to develop and implement increasingly powerful digital minds that no one, not even their creators, can reliably understand, predict, or control.”

The letter is signed by Elon Musk, who originally promoted OpenAI —the company that created ChatGPT— and who, after trying to control it, is now trying to trip it up.

What is the problem?

That the risks that Musk and the other signatories describe are futuristic and science fiction, not the real and pressing ones.

They speak of the danger of creating “non-human minds that could eventually outnumber, outsmart, outdate and replace us” and that we could “lose control of our civilization”.

That will not happen today or tomorrow: we are not at a time when Skynet, the evil intelligence, is going to release a Terminator like in the famous movie.

The biggest threat we have today with artificial intelligence is that its capabilities are concentrating more power, wealth, and resources in a small handful of companies: Google, Microsoft, Facebook, Amazon, and so on.

Precisely, the same companies that are monopolizing all the developments and research in this field,

_

For example: the first signatory to that letter, along with Musk, is Yoshua Bengio, the

of artificial intelligence from the University of Montreal, who sold his deep learning company to Microsoft and went on to become a consultant to the company.

Now, Microsoft has invested 10,000 million dollars in OpenAI, to later integrate the chatbot into its browser.

The most powerful companies on the planet are now engulfing an essential field of research, while the letter warns of a future dystopia in the form of the Terminator.

However, people are suspicious.

A recent study asked more than 5,000 Spaniards about their perception of artificial intelligence and came up with a striking result: fear of these developments stems from suspicion of the economic interests of those who promote them.

To understand each other: the Terminator is not feared, but rather Cyberdyne Systems, the company that in that fiction created the Skynet program without noticing the consequences of opening that Pandora's box.

Weizenbaum developed Eliza for one purpose, but upon contact with humans it became something else.

His intentions did not matter, because people perceived him differently.

Originally, Facebook's motto was "move fast and break things."

When Mark Zuckerberg released Facebook among university students, what was it for?

To remind you of your school friend's birthday, let you flirt with people around you and allow you to share thoughts with the world.

What was the emerging ability?

Collaborate in genocide, as has been confirmed in various parts of the planet.

Why did something like this happen?

Due to the greed of its owners, who already knew the impact on humanity, but also knew that putting the brakes on it hurt their bottom line.

And now, why don't we stop talking about these clever programs, which they had been developing in an opaque way for years?

Because all these companies are in a hurry to make money in the new “race without control” of internet services.

At this point, it doesn't matter what the machine does or how it does it, whether it's a stochastic parrot or a smart one.

What matters is who drives it, why, and what that parrot does to us, regardless of original intentions, as happened with Eliza.

And that is where we should focus with policies that regulate progress, demand transparency and limit the concentration of this new power.

You can follow

Subscribe to continue reading

Read without limits

Keep reading

I'm already a subscriber