If you’ve read the news recently, you’ll see that talk of AI (artificial intelligence) has been increasingly spoken about. What was predicted in cheesy 80s sci-fi flicks seems all the more possible with each passing day. But what is AI and should we be worried about it?
What is AI?
AI has been around much longer than you’d expect, at least since the 1950s. However, the definition of Artificial Intelligence has changed since then. In a nutshell, AI is defined by computers and machines mimicking human problem-solving and decision-making. Voice assistants such as Amazon Alexa, Google Home and Apple’s Siri are all examples of weak AI. This is an artificial intelligence designed to carry out specific tasks based on programming or algorithms.
What sorts of things use AI?
You may have heard of Chat GPT (or other GPTs) which are artificial chatbots capable of holding a conversation with the user. Standing for Generative Pre-trained Transformer, they can answer questions, translate languages and more. Self-driving delivery robots and cars, although not common in the UK, also use AI. Programs like Photoshop now include AI generation which can generate an image based on a prompt. Typing for example, “Mountain landscape in autumn” would produce an image using millions of other images as reference. There’s currently an online debate about the use of these images, as AI uses existing images to learn. Many artists and photographers are debating the use of their media in such training, as they are not given credit. So ultimately, who owns the finished AI-generated image?
The technological world can be quite confusing. Luckily we have a glossary list of words and descriptions below for things that you may have seen appear regularly in the media as of late. Please note, that some of the below are achievable without AI, such as a deep-fake. However, AI can also be used to automatically run these processes so by-hand editing is not required.
So let’s start with our title word. What is a deepfake? According to Google, Deepfake – /ˈdiːpfeɪk/ is a video of a person in which their face or body has been digitally altered so that they appear to be someone else. This can also be used maliciously or to spread false information. You may remember this video by Ctrl Shift Face on Youtube, where Jim Carrey’s face has been digitally applied to Jack Nicholson in the movie The Shining. Although just for entertainment purposes in this instance, the deepfake is so convincing that you could understand why people may think it was real. This is why it’s important to understand when something may be a deepfake, to prevent false information from spreading from sources you think you can trust.
How can I spot a deepfake?
The number one way to tell a video is a deepfake is normally by observing the facial movements. Uncanny or robotic facial movements are normally a dead giveaway that the video may be a deepfake. Issues such as skin-tone changes, and unusual placement of facial features (especially eyes) can make it more obvious that the video may be a deepfake.
Why would a deepfake be dangerous?
Will rapidly evolving AI, you can not only just manipulate videos and people’s faces, but replicate someone’s speech using snippets of audio. Because of this, it’s now possible to make world leaders, celebrities, and family members act and speak exactly how you want. Imagine a video of someone in power spreading online sharing untrue or dangerous information. Some people may automatically trust this source as deepfakes and AI-generated speech can be incredibly convincing.
Virtual Assistant/Smart Assistant
Virtual assistants are internet or cloud-based programs that run set tasks following a prompt. Alexa, Cortana, Google and Bixby are all examples of Assistants that are built into mobile devices or standalone smart hubs. These can do everything from answering questions to turning off your heating. Devices such as lights and plugs that have “smart” on the box usually can be controlled by these types of assistants.
Smart Hubs can be used as a one-stop controller for a lot of your devices. You can simply say “Alexa, turn the living room lights off.” or “Google, let me know the weather for tomorrow.” and your assistant will respond. These are a type of weak AI.
You may have heard talk of Chat GPT recently. Chat GPT is an AI model that allows you to have human-type conversations with the chatbot. It was created by OpenAI who are also responsible for the Dall-E AI art generators. Chat GPT runs on an advanced language model which at the time of writing this, is now up to its fourth iteration. They are trained using a vast amount of media including books, websites, news articles and the conversations it has with the users. Unfortunately, this does mean that on occasion, the chatbot can make a mistake.
Following on from above, Open AI are also responsible for the Dall-E art generator. Dall-E uses the user’s prompt to make visuals in a variety of styles including sketches, oil paintings or realistic photography. There are concerns regarding the generation of images, as the AI uses existing images in order to train. Unfortunately, that means that users can generate images that copy the styles of existing artists. These people may or may not have given their permission for their art to be used to train AI. There is currently no agreement on who can own the copyright of these images, also, would it be the person who generated the prompt? The Artists who trained the AI, or the Dall-E service itself?
Dall-E and other image-generating AI services can also be used to generate images of existing people such as politicians and celebrities, which could potentially be malicious. To take things full circle, these can be used as deepfakes or as part of deepfakes to mislead the public and for malicious intent.
Although AI has already proven itself to be a useful tool, care should be taken when using smart technology or browsing online. This is particularly important when consuming media such as articles and videos. As information can be fake but appear credible at first glance.
We hope we helped clear some of the confusion where AI is concerned. If you’d like to visit us in Warrington for your stairlift needs, you’re welcome at our showroom Monday-Friday. You can also contact us via live chat, on the phone or via email.