The AI Advantage: Exploring the Benefits of Different Models

The AI landscape is rapidly evolving, with new models and applications emerging constantly. This article explores some of the latest advancements in AI, focusing on Apple’s new Intelligence features, a comparison of ChatGPT platforms, and a review of the popular AI chatbots like Claude, Copilot, Gemini and Perplexity. We also delve into the world of AI image generation, examining the strengths of MidJourney and Dall-E.

Apple Intelligence

One of the newest models of AI that was released in the second half of 2024, is Apple Intelligence. This generative AI is designed to improve communication, work and self-expression. It’s supported on more recent devices updated to IOS 18.2, iPadOS 18.2 or MacOS Sequoia 15.2.

A new feature that has upgraded user productivity and communication is the writing tools. Most usable in the Mail app, Apple’s writing tools can summarise long emails and messages into short and concise versions that only highlight the important points. Another rather handy feature to note, is the possibility to have the intelligence proofread or rewrite emails varying in wording and tone. For example, it can rephrase your writing to make it sound more friendly, professional, or succinct. These tools can also be used in the Messages, Notes and Pages app. The only downside is these tools are not available to use on a different email platform such as Outlook or Gmail (even on your compatible Apple device).

With the update, Apple has also introduced a new app called Image Playground. This allows users to experiment with AI to create original illustrations based on descriptions, concepts and even people from their Photo library.

Also, Siri now has access to ChatGPT. This means that when you ask a question, Siri is able to give you answers sourced from ChatGPT, enhancing Siri’s capabilities and answers. However, users can control this if they do not wish to use this feature.

ChatGPT

ChatGPT is a popular AI chatbot which is often utilised for answering questions, creating images, generating text and gathering information. GPT-3 is free to access but is limited in its use. On the other hand, GPT-4 requires a paid subscription. GPT-3 works great for general use such as researching a topic, writing an article or answering basic questions. The most noticeable difference in the paid version is that the accuracy, creativity and capabilities tend to be better, than those of the free version. GPT-4 tends to give more detailed responses and is less likely to make factual errors. Furthermore, when prompted to generate long text, GPT-4 shows greater creativity and consistency, which in comparison, GPT-3 can start to produce repetitive and sometimes incoherent responses. Overall, GPT-3 is more useful for assistance, whereas GPT-4 is worth purchasing and is better to use if you are more heavily relying on the information and responses.

Claude

Anthropic’s Claude is another AI chatbot used to generate human-like responses and perform tasks such as creating content, summarising information and answering questions. A helpful feature of this model is the option to choose a response style such as normal, concise, formal and explanatory. There is also the possibility to create your own style by customising details like tone, voice, vocabulary etc. or by uploading your own example of a writing style for Claude to personalise its responses with, making it a great tool for generating written pieces.

Copilot & Gemini

Microsoft’s Copilot and Google’s Gemini are two more AI models. Although these chatbots provide a similar function, Copilot is ideal for those who depend on Office 365, whereas Gemini works better for those who use Google Workspace and other Google services. You’ll find that Copilot has been integrated into Microsoft Word, Excel, PowerPoint, Outlook and Teams, to assist users with tasks such as creating drafts, analysing spreadsheets, generating presentations and summarising emails. Similarly to Copilot, Gemini has also been integrated into Google’s own ecosystem of Docs, Sheets, Maps and Gmail etc. As a recent application of Gemini, Google users may have noticed an AI Overview pop up when googling something. This feature appears at the top of a search result, using generative AI to provide a quick answer to the prompt. The summary offers a snapshot of the topic with key information and links to other relevant sources, to supply the user with a brief but insightful understanding of their search.

Perplexity AI

Like many of the other conversational AI models, Perplexity AI utilises natural language processing to provide direct and informative answers to user queries. One of the major benefits to using Perplexity is how the platform typically cites its information sources. This allows users to verify the credibility of the responses, and in addition, click on the links supplied to check out the original page from where the information was gathered. Furthermore, Perplexity will also source relevant images, videos and recommend other related topics to enhance the user’s search, which can be particularly helpful to those conducting research.

Image AI

MidJourney and Dall-E are image AI platforms, developed for generating images based on text descriptions. MidJourney is well known for its incredible visuals, surreal aesthetics and creativity, making it popular amongst artists and designers. It does however require a paid subscription to use. Dall-E is included in GPT-4. The images generated are usually quite accurate to the prompts and also permit users to modify specific parts of an image at request. Dall-E is useful in the conceptualisation of people, ideas and objects. Comparing the two, MidJourney can create a much wider range of artistic styles. MidJourney’s accuracy and detail tends to outperform Dall-E in creating realistic images, overall producing higher-quality results.