ChatGPT and other AI-powered apps to keep an eye on over the next few months

One thing is certain: artificial intelligence (AI) is a technology that will change the way we live and work most in the near future. With the launch of ChatGPT – the AI-powered chatbot that seems to have a convincing answer to any question – interest in AI and its applications is exploding.

Everyone from computer scientists and investors to businesses and first adopters is on the look out for the next big thing in Ai-powered apps.

AI platforms and apps don’t just promise to keep us company and answer our questions, but to help us in many other ways too: they will improve the online search experience, help companies provide customer care, enable us to write and polish articles, suggest lines of code and advise on how to increase business productivity.

So what AI-powered apps are currently on the market? What is the technology behind them? And what do they do exactly? We take a look at the products already out there and others that are in the pipeline.

Here’s our pick AI systems, apps and platforms to watch over the next few months.

• ChatGPT – Currently the world’s most talked about chatbot, it’s powered by conversational AI
• Google Bard – Google’s answer to ChatGPT, designed to revolutionise the search engine world
• Midjourney – an AI program for creating images from text prompts
• DALL·E 2 – another AI-powered text-to-image system to watch closely
• Writesonic – an AI-powered tool for writing articles and content for social media
• Synthesia – an AI platform for creating videos
• GitHub Copilot – an app that helps developers to write source code for software using AI

ChatGPT

Where else to start, if not here? ChatGPT is the AI app that people have been talking about for months. The announcement of its launch in November 2022 is already considered by many to be a key event in the history of technology, right up there with the unveiling of the iPhone in 2007.

Technically, ChatGPT is a chatbot, in other words a conversational AI app. It understands natural language and responds intelligibly to questions asked by users in that language. You can ask ChatGPT pretty much anything and in a completely natural way, like: “Can you translate this text for me?” or “What is the French revolution?” Every day, ChatGPT’s more than 600 million monthly users ask it to talk about everything from poetry to coding, from academic essays, to school homework and philosophical questions.

Currently, ChatGPT is based on the GPT-3.5 model, an artificial neural network that has been trained on hundreds of billions of online texts. The app was developed by OpenAI, a company founded by Elon Musk and Sam Altman and heavily backed by Microsoft.

Of course, at the time of writing (March 2023) ChatGPT still has limitations. The training database stops at 2021: this means that ChatGPT doesn’t know what has happened over the past two years. Odd, right? What’s more, it’s text-based only: it can neither understand nor generate images, videos or voice prompts, although some of these obstacles will be overcome with the update to the GPT-4 model.

At the moment, the programme is free to access online here if you sign up, but because of the high volume of queries, it’s often hard to use.

In short, ChatGPT:

• Is an app powered by conversational artificial intelligence (a chatbot)
• Responds to text inputs providing answers in natural language
• Is text-based only; it doesn’t understand audio or images
• Currently free

Google Bard

Google Bard was unveiled as Google’s answer to ChatGPT. At the moment (March 2023), this AI app has not yet been made public and is only accessible to a very small number of users.

So far, we know that Google Bard, like ChatGPT, is a conversational AI app, more commonly known as a chatbot. Although we still don’t have precise details about what it can do, Google Bard should be able to respond coherently to different types of user input in the form of a conversation. Google is working to integrate Bard into its search engine in an attempt to revolutionise the search experience by making it more reliable and closer to our own language. Likewise, ChatGPT is about to be integrated into Bing, the biggest rival search engine to Google.

Google Bard is built on LaMDA (Language Model for Dialogue Applications), a family large language models that Google has been developing since 2020.

In short, Google Bard:

• Is a chatbot similar to ChatGPT but developed by Google
• Was unveiled in February 2023, but is not yet public
• Will improve the search experience on Google

Midjourney

Put simply, Midjourney does for images what ChatGPT does for text.

Midjourney is a text-to-image AI app, in other words, an AI system that can create images from text inputs. A user could, for example, ask it to create illustration of a woman on a bicycle in the style of Picasso or a cyber-punk mountain landscape.

Once it has received the first prompt (the technical term for the text input by the user), the app outputs 4 images ready for further editing to obtain the desired result: for example, you could ask the program to remove buildings from the background, use a particular format or adopt a totally different style.

Midjourney was launched in beta version (which remains the version currently available) in July 2022 by a small independent lab based in San Francisco. The system uses machine learning and was trained on an immense database of digital images.

From the very outset, it has attracted interest and criticism in equal measure. The cover of the 11 June 2022 edition of The Economist was created with Midjourney, an issue that explored the legal, ethical and philosophical questions raised by the use of artificial intelligence.

You can get a free trial of Midjourney or use it with a paid licence. You can access Midjourney here.

In short, Midjourney:

• An AI app that creates images from text prompts
• Lets you edit these images by adding details, styles or technical parameters
• Is still in beta version but can be used with a paid plan or the free version

DALL·E 2

Like Midjourney, DALL·E 2 is an AI-powered text-to-image app for creating images.

Launched in 2021 by OpenAI, the same company behind ChatGPT, the app is already on its second, more powerful version. The neural network that powers DALL·E 2 was trained on a massive database of images accompanied by a text description.

DALL·E 2 uses the same technology as ChatGPT, the only difference being that it interprets natural language to generate corresponding images. For example, in response to the prompt “a gorilla playing on the computer” the app will generate a realistic image. Not only can the app create specific images from scratch, it can also edit them in response to simple text prompts: if you don’t like the gorilla, you can ask the app to “replace the gorilla with a teenager”.

At present, Midjourney and DALL·E 2 are the two best-known examples of text-to-image AI: it will be interesting to see how they develop and differ. For now, the main differences lie in style, with DALL·E 2 tending to produce images that are more photographic than Midjourney.

DALL·E 2 can be accessed here. You can create your first few images for free and then buy additional credits to increase your output.

In short, DALL·E 2:

• Is the leading text-to-image AI app, together with Midjourney.
• Tends towards a more realistic and photographic style compared to competitors
• Was launched by OpenAI, the firm behind ChatGPT

Writesonic

Writesonic is an AI-powered app for creating and optimising text and articles: it’s a sort of digital copywriter that uses an AI engine.

Using natural language processing models and machine learning, Writesonic can create fairly complex articles, content for social media and copy for online advertisements. The app can also use AI to improve texts that have already been written, for example, by optimising them for search engines.

The first fairly rudimentary version of Writesonic was launched in 2020 by its creator, the Indian computer scientist Garg Samanyou, with the aim of combining the power of artificial intelligence with human creativity. Since then, Writesonic has also created its own alternative to ChatGPT, ChatSonic.

Like ChatGPT, ChatSonic is an app powered by conversational AI that responds to text inputs. Unlike its better known rival, ChatSonic is also able to generate images and even understand voice commands. The has been trained on an up-to-date database, unlike ChatGPT’s database, which stops in 2021.

You can try Writesonic for free up to a word limit, beyond which you have to buy a paid plan. You can access this app here.

In short, Writesonic:

• Is an AI-powered app for creating and optimising articles
• Can be used to create or optimise articles, blog posts, content for social media and advertisements
• Has a sister app called ChatSonic, an alternative chatbot to ChatGPT

Synthesia

Synthesia is an AI-powered video creation platform. The app was launched in 2017 by a group of researchers from Stanford, UCL, TUM and Cambridge universities.

Synthesia is browser based: when you type in a text input, a presentation video is created in which an avatar reads out the text provided by the user. The advantage it offers businesses is the ability to create lots of videos without having dedicate an entire team to the task and use expensive equipment like microphones and cameras.

The app uses an AI engine that adapts the movement of the avatar’s lips to what it is saying and uses a very realistic synthesised voice (here’s an example of a promotional video for a finance app).

Synthesia currently lets you choose between over 85 avatars based on real actors, as well as more than 120 languages. In some languages, you can also change the accent. To access a demo version or get a quote for your needs, you have to contact the company. The basic version costs £23 a month.

In short, Synthesia:

• Is an AI-powered app for creating promo videos
• Is an AI power-app that synthesises natural voices and synchronises speech with the lips of its avatars
• To request a demo version, you need to contact the company

GitHub Copilot

Did you know that AI-powered apps  can also help developers in their work? GitHub Copilot, for example, is a platform that assists developers in writing code: it’s a sort of developers’ assistant.

When writing code, the app suggests various ways in which the code could be completed in order to speed up the process and lighten the workload. GitHub Copilot is based on GPT-3, the same technology behind ChatGPT: however, in this instance the AI model has been specifically designed to understand and generate both natural and programming languages.

At the moment, it doesn’t seem capable of doing anything more than play the role of a simple assistant and can’t handle a complex system on its own. But in the future, who knows?

You can find GitHub Copilot here.

In short, GitHub Copilot:

• Is an AI-powered assistant for for writing code
• Has an AI engine understands and generates both natural languages and programming languages
• Does not seem to be able to handle complex systems on its own.

What about you? Have you tried any AI-powered apps yet? What are your first impressions?