Quantcast
Channel: Phil Simon
Viewing all articles
Browse latest Browse all 25

The Dangerous Democratization of AI

$
0
0

Image from Craiyon

In 2004, one of the world’s smartest techies made what seemed like a reasonable prediction. He promised that the powers that be would “solve” spam within two years.

That luminary was, of course, Bill Gates of Microsoft fame. You probably heard of him.

Last year, spam remained alive and well. The arrival of generative AI means many things. As I wrote in The Nine, unsolicited scams are about to become much more difficult to identify.

It was only a matter of time before ChatGPT faced stiff competition—and not just from Google, Meta, Amazon, and other tech giants. Open-source large-language models have arrived in earnest, as Ethan Mollick writes in his latest Substack:

since anyone can modify these systems, to a large extent AI development is also now much more democratized, for better or worse. For example, I would expect to see a lot more targeted spam messages coming your way soon, given the evidence that GPT-3.5 level models works [sic] well for sending deeply personalized fake messages that people want to click on.

So long, obvious Nigerian diamond-mining schemes. More people will fall for these increasingly plausible ruses, and the consequences should terrify us. Tricks can help, but nothing is failsafe.

Simon Says

I’ve said it dozens of times in the past year: You can ignore any or all of the forces in The Nine, but they sure as hell are not ignoring you. Foolish is the soul who pretends otherwise. Blockchain technologies will help combat spam, but your best weapon will continue to be that old-school organ between your ears.

Image from Craiyon

In 2004, one of the world’s smartest techies made what seemed like a reasonable prediction. He promised that the powers that be would “solve” spam within two years.

That luminary was, of course, Bill Gates of Microsoft fame. You probably heard of him.

Last year, spam remained alive and well. The arrival of generative AI means many things. As I wrote in The Nine, unsolicited scams are about to become much more difficult to identify.

It was only a matter of time before ChatGPT faced stiff competition—and not just from Google, Meta, Amazon, and other tech giants. Open-source large-language models have arrived in earnest, as Ethan Mollick writes in his latest Substack:

since anyone can modify these systems, to a large extent AI development is also now much more democratized, for better or worse. For example, I would expect to see a lot more targeted spam messages coming your way soon, given the evidence that GPT-3.5 level models works [sic] well for sending deeply personalized fake messages that people want to click on.

So long, obvious Nigerian diamond-mining schemes. More people will fall for these increasingly plausible ruses, and the consequences should terrify us. Tricks can help, but nothing is failsafe.

Simon Says

I’ve said it dozens of times in the past year: You can ignore any or all of the forces in The Nine, but they sure as hell are not ignoring you. Foolish is the soul who pretends otherwise. Blockchain technologies will help combat spam, but your best weapon will continue to be that old-school organ between your ears.

The post The Dangerous Democratization of AI appeared first on Phil Simon.


Viewing all articles
Browse latest Browse all 25

Trending Articles