We have to stop ignoring the problem of AI deception.

WhatsApp Group Join Now
Telegram Group Join Now
Instagram Group Join Now

Google I/O introduced an AI assistant that can see and hear the world, while OpenAI introduced its own version his– Similar to chatbot in iPhone. Next week, Microsoft will host Build, where it's sure to have some version of Copilot or Cortana that understands PivotTables. Then, a few weeks after that, Apple will host its developer conference, and if the buzz is anything to go by, it'll be talking about artificial intelligence. (Not sure if Siri will be mentioned.)

AI is here! It is no longer imaginary. It's hiring, creating some new jobs, and helping millions of students avoid doing their homework. According to most of the big tech companies investing in AI, we Technology appears to be about to experience one of those rare monumental changes. Think of the Industrial Revolution or the creation of the Internet or the personal computer. All of Big Tech's Silicon Valley—centered on taking big language models and other forms of artificial intelligence and moving them from researchers' laptops to average people's phones and computers. Ideally, they will make a lot of money in the process.

But I really couldn't care less because the meta AI thinks I have a beard.

I want to be very clear: I am a cis woman and do not have a beard. But if I type “show me a picture of Alex Crane” into the prompt window, MetaAI inevitably returns pictures of very handsome dark-haired men with beards. I am just some of these things!

Meta AI isn't the only one to struggle with. the edgeThe masthead of ChatGPT told me yesterday that I don't work the edge. Gemini from Google didn't know who I was (fair) but after telling me Nelle Patel was the founder. the edge, He then apologized and corrected himself saying he wasn't. (I assure you he was.)

AI keeps getting worse because these computers are stupid. Extraordinary in their abilities and astonishing in their dullness. I can't get excited about the next turn in the AI ​​revolution because that turn is one where computers can't be consistently accurate about even trivial things.

I mean, they even messed up during Google's big AI keynote at I/O. In a commercial for Google's new AI-ified search engine, someone asked how to fix a jammed film camera, and he suggested “open the back door and gently remove the film.” This is the easiest way to destroy any photo you've taken.

Some of these tips are good! Some require a very dark room.
Screenshot: Google

AI's difficult relationship with reality is called “hallucinating.” Put very simply: these machines are very good at discovering patterns of information, but in trying to extract and create them, they sometimes get it wrong. They effectively “fake” a new reality, and that new reality is often false. It's a tough problem, and everyone working on AI right now knows it.

A former Google researcher claimed it could be fixed within the next year (though he lamented the conclusion), and Microsoft has a tool for some of its users to help them detect it. Provides assistance. Google's head of search Liz Reed said the edge It is also aware of the challenge. He told my colleague David Pearce that with any language model “there is a balance between creativity and reality.” “We're really making it a reality.”

But how was the balance, said Reid? That's because many AI researchers aren't really delusional. May be Solution A study by the National University of Singapore suggested that hallucinations are an inevitable consequence of all major language models. Just as no one is 100 percent accurate all the time, neither are computers.

And that's probably why most of the big players in the field—those with the real resources and financial incentive to push us all to adopt AI—think you shouldn't worry about it. During Google's IO keynote, it added the phrase “check answers for accuracy” to the screen below nearly every new AI tool, in small gray font — a helpful reminder that its tools are not to be trusted. Could be, but don't think it's a problem either. ChatGPT works the same way. In small font at the very bottom of the prompt window, it says, “ChatGPT may make errors. Check important information.”

If you look closely, you can see the small and slanted opening.
Screenshot: Google

This is not a withdrawal you want to see from the tools that are going to change our entire lives in the near future! And the people who make these tools don't care much about fixing the problem other than a small warning.

OpenAI CEO Sam Altman, who was briefly ousted for prioritizing profit over safety, went a step further and said that anyone who had a problem with the accuracy of AI was naive. “If you just act naive and say, 'Never say anything you're not 100 percent sure of,' you can get them all to do it. But in this It won't be the magic that people love so much,” he told a crowd at Salesforce's Dreamforce conference last year.

The idea that AI has some kind of inauthentic magic sauce that will allow us to excuse its tenuous relationship to reality has been taken up a lot by people keen to address validity concerns. Google, OpenAI, Microsoft, and many other AI developers and researchers have dismissed illusion as a minor annoyance that should be forgiven because they are on the way to creating digital beings that can transform our own lives. Can make it easier.

But apologies to Sam and everyone else who financially incentivized me to get excited about AI. I don't come for the false magic of human consciousness on computers. I come to them because they are so accurate when humans are not. I don't need my computer to be my friend. I need it to correct my gender when I ask and to help me avoid accidentally exposing film when fixing a shuttered camera. I think lawyers would want the case law to be correct.

I to understand Where Sam Altman and other AI evangelists are coming from. In some distant future these and Xero are likely to create true digital consciousness. At the moment, the development of artificial intelligence is moving at a staggering pace that puts many previous technological revolutions to shame. Real magic is at work in Silicon Valley right now.

But the AI ​​understands that I have a beard. It can't consistently figure out the simplest of tasks, and yet, it's being foisted upon us with the expectation that we celebrate the incredible mediocrity of the services these AIs provide. While I can certainly marvel at the technological innovations taking place, I'd rather my computers not sacrifice accuracy, so I have a digital avatar to talk about. It's not a fair trade-off – it's just an interesting one.

WhatsApp Group Join Now
Telegram Group Join Now
Instagram Group Join Now

Leave a Comment