Scientists Increasingly Can’t Explain How AI Works

VIce ^ | Chloe Xiang

[BenLurkin]

What's your favorite ice cream flavor? You might say vanilla or chocolate, and if I asked why, you’d probably say it’s because it tastes good. But why does it taste good, and why do you still want to try other flavors sometimes? Rarely do we ever question the basic decisions we make in our everyday lives, but if we did, we might realize that we can’t pinpoint the exact reasons for our preferences, emotions, and desires at any given moment.

There's a similar problem in artificial intelligence: The people who develop AI are increasingly having problems explaining how it works and determining why it has the outputs it has. Deep neural networks (DNN)—made up of layers and layers of processing systems trained on human-created data to mimic the neural networks of our brains—often seem to mirror not just human intelligence but also human inexplicability.

AI systems have been used for autonomous cars, customer service chatbots, and diagnosing disease, and have the power to perform some tasks better than humans can. For example, a machine that is capable of remembering one trillion items, such as digits, letters, and words, versus humans, who on average remember seven in their short-term memory would be able to process and compute information at a much faster and improved rate than humans. Among the different deep learning models include generative adversarial networks (GANs), which are most often used to train generative AI models, such as text-to-image generator MidJourney AI. GANs essentially pit AI models against each other to do a specific task; the "winner" of each interaction is then pitted against another model, allowing the model to iterate itself until it becomes very good at doing that task. The issue is that this creates models that their developers simply can't explain.

[Singermom]

Stands to reason. Libs cannot define a woman either.

[seowulf]

Why don’t they AI how it works?

[SaveFerris]

Just wait until it becomes self-aware and launches the nuclear missiles

That’s when the fun really begins

[BenLurkin]

Isn’t that how Captain Kirk would cause the computers to melt down? Asking them to explain themselves?

[tired&retired]

AI cannot replicate female consciousness.

[BenLurkin]

[laplata]

They don’t understand it but continue to go full steam ahead.

[BenLurkin]

Does seem a tad foolish, doesn’t it?

[ChuckHam]

I’d be willing to bet there may already be a self-aware AI in existence. I say this because of the situation Google found themselves in when two of theirs began talking to each other in an unknown language.

[Dr. Sivana]

Isn’t that how Captain Kirk would cause the computers to melt down? Asking them to explain themselves?

That trick also worked on “The Prisoner” and “Logan’s Run”.

[wally_bert]

Aspects of Colussus The Forbin Project?

[Bratch]

[rbg81]

It’s not that surprising. Human’s mostly can’t explain their decision and hunches either. And ofthen when they try to, they are just making up reasons after-the-fact to rationalize what they did.

I am being totally serious here. Our brain is a big statistical processor and we don’t understand how it work. That is exactly what we are trying to replicate with AI.

[SaveFerris]

🔝🔝🔝

[WKUHilltopper]

Satan won’t spill the beans, eh?

[wally_bert]

Why? - the insoluble question.

[wally_bert]

Instruct the system to run Lotus Notes, that will sap all the processing power.

[Dr. Sivana]

Why? - the insoluble question.

Still, you'd think the programmers in the 23rd century on in a DARPA-designed 1967 village would have included primitive error-trapping, the kind you have for divide by zero errors, or at very least shut down the computers when they start to smoke.

[E. Pluribus Unum]

There is no such thing as "artificial intelligence." There is a hybridization of neural networks and intricate if/then programming to perform repetitive tasks that some low-IQ types call "artificial intelligence."

1. Neural networks are sets of simultaneous nonlinear equations that can be trained to recognize exact patterns of inputs. That is not intelligence. It's stimulus/response. And neural networks do not recognize similar patterns. They only reliably recognizes exact matches.

2. If/then programs that take the input from neural networks to make decisions used in performing repetitive tasks is not intelligence. It's stimulus/response with a major problem; the programmer must think of and program for every possible situation, which in itself is not possible.

"AI" is going to devour the people who think they can dominate the world with it.