April 15, 2024

10 things you need to know about AI

From Google Maps to facial recognition, we are increasingly reliant on artificial intelligence. Expert MIKE WOOLRIDGE gives his reasons for being afraid and being happy.

We asked ChatGPT (chat.openai.com) to create a lead image for this article based on the following prompts from YOU: “the future looks bright”, “flat illustration”, “poster style”, “embrace technology in our favor” and “older woman’

1 AI is not that smart as it seems (still)

The only golden message is that even with the next generation of conversational software tools like ChatGPT, which is the biggest thing (possibly ever) in AI, it’s not about a mind. ChatGPT [an AI-powered large language model that is trained on huge amounts of data] is a computer program that has been highly optimized, to use the technical term, to do one thing: tell you what you want to hear.

He’s very good at it: very fluid and very eloquent. That leads us to believe that we are communicating with a mind, like a human being, but that is not the case.

2 Can’t cope with the real world

AI is about getting computers to do things that only the human brain and nervous system can do. AI finds some things that we don’t associate with intelligence, like driving a car, incredibly difficult. Untold amounts of money have been spent trying to get cars that can drive themselves safely and we’re not there yet. We are a long way from having robotic butlers.

3 Google Translate It’s a game changer

AI translation tools are one of the most amazing things humanity has achieved. You can translate practically any language used in the world. Future AI translation applications will learn about you, the language you use, and the way you communicate. You will be able to have simultaneous translations of spoken words.

4 Don’t trust ChatGPT with your credit card

One of the fundamental problems with large language model (LLM) technology is that it doesn’t tell the truth. You ask him a question; gives you the most plausible answer. He has no conception of truth. If you don’t know the answer to a question, you’ll just make something up. I asked an older version of ChatGPT what he knew about me. His two-line summary said I had studied at Cambridge, which is the typical training for an Oxford don, but I haven’t. That’s why I would be reluctant to put one of these LLMs in a situation where they had to make decisions with real consequences. The technology is unreliable and prone to strange errors. I wouldn’t trust my credit card to ChatGPT.

5 Don’t believe everything you read on social networks

AI can access your feed, look at your personal biases, and produce industrial-scale misinformation about your favorite politician—or least favorite, depending on the language you’ve used online—so you can relate to them if you’re a teenager. in Middlesbrough or a Conservative voter in the home counties. It is easy to see how this could be used to influence voters and affect political elections in the UK and abroad.

Within a few years, it is very likely that there will be much more AI-generated content on social media than human-generated content. It will be very difficult for us to know what is authentically human and what is not. Trusted news sources will be extremely important in navigating that. Fast forward 50 years, and there will be many, many more AI-generated books, music, and videos than human-generated ones.

6 AI will transform healthcare

Doctors will use AI as a diagnostic tool in the same way that using an X-ray machine is a tool. But I would caution against relying entirely on an AI diagnosis. My big concern is that when a machine tells you something that doesn’t agree with your personal perceptions, you have to invest some energy into figuring out why the machine might be wrong. And a lot of times people don’t bother doing that if they’re tired. I would rather have a human than an AI. Some people won’t have that luxury: it will be AI or nothing. On the positive side, artificial intelligence tools will make very expensive healthcare, only accessible to a small fraction of the world’s population, accessible to large numbers of people. For example, AI technology will be able to analyze scans from the other side of the world. That will be transformative.

7 It is biased towards certain sectors of society.

AI is dominated by white, college-educated, American men. That’s a widely recognized problem: LLMs communicate using North American (rather than British) English. This matters. A recent widely used data set for speech recognition technology used recordings made by white, male, college-educated Americans. If you were outside that data set (a woman, for example), voice recognition systems didn’t do a good job of understanding you. To create an LLM, you need many examples of the language. Icelandic does not currently have a large enough digital footprint to create LLMs that can communicate in that language; that’s a subtle danger about the way AI may not encompass a broader range of human viewpoints and experiences.

8 There is a risk of misuse

Most human programmers would reject a request to write a program to launch a cyber attack on the NHS data servers. But if you have an LLM that will do everything you ask, attacks like that are easier. [in the wrong hands]. People using AI to do bad things is something we need to think about. At a recent AI summit in the UK, the debate moved from existential risk to catastrophic risk, whereby an AI failure leads to something (such as a plane crash in a densely populated area) that would be devastating, but not will lead to the end of humanity. That is a healthy direction to take in the discussion, because it is more real, and when it is more real it is something that you can think about mitigating.

9 AI is just a tool

AI is not something to be amazed or feared. It will be used to do wonderful things, like recognize tumors on x-ray scans and detect dementia or early-onset heart disease, but it is just a tool like any other tool; Don’t talk about it as if it were a person.

10 Could Boost Creativity

The goal is for AI to free people from drudgery and allow them to focus on important things that require human understanding, creativity and empathy. An unexpected application of LLMs is brainstorming new ideas – something that people don’t find easy or value. Ask ChatGPT for offers on a banana-flavored dairy drink and it will keep churning out, trash or not. And they will get the human creative juices flowing.

LLMs will open up opportunities for completely new art forms that we can’t even imagine at the moment, just as whoever invented the first synthesizers couldn’t have imagined the take-off of rave music in the 80s. I think we’re on the cusp of an explosion. of people doing really creative things with AI.

Mike Wooldridge is a professor of AI at the University of Oxford. The truth about AIhis Royal Institution Christmas lectures are on iPlayer.

Illustration: Chloe Sharp/Midjourney

Leave a Reply

Your email address will not be published. Required fields are marked *