Intro to AI and the Major Players in 2025

By Scot Hulshizer | January 20, 2025

Artificial Intelligence (AI) is one of the most transformative technologies of our time, reshaping industries and influencing daily life. From chatbots to multimodal systems capable of processing text, images, and audio, AI continues to evolve rapidly.

In this article we will explore the fundamentals of AI, its rapid development, its impact on jobs, and a detailed look at ten of the most prominent AI models popular in North America.

We'll also talk about how we help small and mid-market business owners leverage this technology in an affordable and responsible way.

What is AI (artificial intelligence)?

Simply put, AI is about creating machines or systems that can act like humans by learning from information, solving problems, and making choices.

These systems use algorithms (a process or set of rules to be followed in calculations or other problem-solving operations) and lots of data to do things like understand language, recognize images, and make decisions.

Types of artificial intelligence

AI can be classified in a few different categories based on its capabilities or its functionality:

AI classified by capabilities:

Narrow AI (Weak AI) - This type of AI is good at specific tasks like translating languages or recognizing faces. Think of Alexa or recommendations from Netflix.
General AI (Strong AI) - This is the kind of AI that could do anything a human can do intellectually. It's still something researchers are working towards.
Superintelligence - This is a theoretical type of AI that would be smarter than humans in every way. It raises questions about control and safety. As rapidly as AI is advancing, outside of Sci-Fi movies we're still quite a way off from this type of artificial intelligence.

AI classified by functionality

Reactive Machines - Basic systems that respond to inputs but lack memory (e.g., IBM’s Deep Blue).
Limited Memory - Systems that learn from past data for better decision-making (e.g., self-driving cars).
Theory of Mind - Emerging AI aiming to understand emotions and intentions.
Self-Aware AI - A theoretical stage where machines possess human-like consciousness.

The strengths and weaknesses of AI

It is easy to get caught up in the idea that AI is the solution to every business problem or the fire-breathing dragon that is about to decimate the workforce and render humans irrelevant.

While certainly the most transformational technology of our time, it still has its place.

Where AI excels

AI has proven itself invaluable in certain areas:

Data Processing: Rapidly analyzing massive datasets for insights (e.g., fraud detection in finance).
Automation: Performing repetitive tasks with speed and precision (e.g., manufacturing robots).
Predictive Analytics: Forecasting trends in industries or financial markets.
Natural Language Processing: Enabling chatbots and translation tools to interact with humans.
Medical Diagnostics: Identifying diseases through imaging analysis with high accuracy.

Where AI falls short

Despite its advancements, AI has notable limitations:

Lack of Generalization: Struggles to apply knowledge across different contexts.
Ethical Concerns: Biases in training data can perpetuate discrimination.
Common Sense Reasoning: Fails at understanding abstract concepts or nuanced language.
Creativity: Limited ability to generate truly novel ideas or solutions.
Emotional Intelligence: Cannot empathize or form genuine human connections.

How quickly is AI advancing?

AI is moving forward incredibly fast. In just five years, we've seen major developments like OpenAI's GPT-4o, Google's Gemini models, and Meta's LLaMA series. These models are bigger and better than ever, able to work with text, images, audio, and even video all at the same time.

This rapid progress is because of:

Increased Computational Power: Companies like NVIDIA and Google are making specialized hardware for training huge AI models.
Data Availability: The internet provides massive amounts of data for training AI.
Investment: Companies like OpenAI, Microsoft, Meta, and Google have invested billions into AI research and infrastructure.

Experts think that by 2030, AI could add $15.7 trillion to the global economy and automate up to 30% of jobs in developed countries.

The impact of AI on jobs

Speaking of jobs, at Acuity Advisor Group we are acutely aware of this impact as it impacts multiple pieces of our practice.

From a business perspective, we are always looking for ways to help owners and executives make better use of technology to improve efficiency and cut costs.

Through our career and executive coaching services, we encounter many knowledge workers who worry how AI will impact not only their immediate job prospects but their chosen profession over time.

Job displacement vs. job creation

Studies suggest that up to 800 million jobs could be replaced globally by AI as soon as 2030. Jobs like customer service agents, writers, and financial analysts are at risk because of advancements in AI that can generate content.

However, AI is also creating new jobs in fields like machine learning engineering, data analysis, and AI ethics. By taking care of boring tasks, AI can free up workers to focus on creative and strategic work.

In good news for those that are willing to pivot, the World Economic Forum predicts that while 75 million jobs may be lost by 2025, 133 million new jobs could be created.

The need for "reskilling"

As industries change because of AI, it's crucial for workers to learn new skills. Areas like programming, data analysis, and critical thinking will likely be the beneficiaries of a growth in artificial intelligence.

White-collar professionals in particular should stay plugged in to current trends and begin building new skills to stay competitive in the future.

How many AI models exist today?

As of 2025, there are over 80 large-scale language models globally, alongside thousands of smaller generative models used for tasks like content creation, code generation, and image synthesis. (We know this because AI tells us so.)

While most focus on narrow applications, large-scale large language models (LLMs) dominate the field due to their versatility and advanced reasoning capabilities.

The Top AI Models in North America

The following are a fairly arbitrary list of the "top" AI models in use in North America. We intentionally excluded a number of fairly significant Chinese models as they are not presently focused on English language processing.

Our selection criteria was largely driven by the scope of implementation and the differences in capabilities.

We have organized the list to keep different solutions by the same provider together.

ChatGPT (GPT-4o) by OpenAI

ChatGPT is powered by OpenAI’s GPT-4o model, an advanced multimodal system capable of processing text, images, and audio simultaneously. Released in December 2024 as an upgrade to GPT-4 Turbo, GPT-4o is faster and more efficient while offering real-time interactions with minimal latency. It boasts over 1.76 trillion parameters, making it one of the most powerful language models ever created.

ChatGPT is widely used for tasks such as writing emails, generating code snippets, tutoring students in subjects like math or science, and analyzing visual data like charts or graphs. Businesses use it to automate customer support, create marketing content, and even analyze legal documents. Its versatility has made it a go-to tool for professionals and casual users alike.

In short, ChatGPT helps people do everything from writing better emails to solving hard problems quickly.

DALL-E by OpenAI

DALL-E is OpenAI’s image-generation model that turns text descriptions into stunning visuals. First launched in January 2021 with subsequent updates (e.g., DALL-E 3), this model has revolutionized creative industries by enabling users to create photorealistic images from simple prompts.

DALL-E is widely used for marketing campaigns, product design mockups, educational materials requiring visual aids, and even fine art creation. The latest version includes built-in safeguards to prevent misuse while offering seamless integration with other OpenAI tools like ChatGPT.

In short, DALL-E converts words into pictures.

Whisper by OpenAI

Whisper is OpenAI’s automatic speech recognition (ASR) system designed for transcribing spoken language into text with near-human accuracy. Trained on over 680K hours of multilingual data from diverse sources like podcasts and interviews, Whisper excels at handling noisy environments or diverse accents.

Whisper is commonly used for live transcription services during meetings or translating audio recordings into English text for global audiences. It has become a vital tool for journalists, podcasters, educators, and businesses conducting international operations.

In short, Whisper turns talking into writing.

Microsoft Copilot

Microsoft Copilot is a generative AI tool deeply integrated into the Microsoft 365 ecosystem. Released in early 2023, it uses OpenAI’s (Microsoft is a major investor) GPT-4o alongside Microsoft’s proprietary Prometheus model to enhance productivity within applications like Word, Excel, PowerPoint, Outlook, and Teams. By connecting to Microsoft Graph, Copilot leverages organizational data such as emails, chats, documents, and meetings to provide contextually relevant assistance.

Copilot excels at automating repetitive tasks: drafting documents in Word, analyzing datasets in Excel, creating presentations in PowerPoint, summarizing meetings in Teams, and drafting emails in Outlook. It also supports real-time transcription and image generation through DALL-E 3 integration. Businesses rely on Copilot to save time and improve efficiency across their workflows.

In short, Microsoft Copilot helps people work faster inside apps they already use every day.

Google Gemini

Google Gemini debuted in December 2023 as Google’s most advanced multimodal model yet. Developed by Google DeepMind as a successor to PaLM 2, Gemini can process text, images, video, audio, and code simultaneously. Its integration into Google Workspace tools like Docs and Sheets has made it an essential productivity tool for millions of users.

Gemini powers Google’s Bard chatbot while also enhancing search functionalities with advanced reasoning capabilities. Its ability to summarize lengthy documents or generate detailed reports makes it invaluable for enterprises seeking efficiency gains.

In short, Gemini makes Google apps more powerful from note-taking to research to polishing up an email.

LLaMA by Meta

Meta’s (owner of Facebook and Instagram) LLaMA series began in February 2023 as an open-source alternative to proprietary LLMs like GPT-4o. The latest version—LLaMA 3—features up to 405 billion parameters with long-context understanding across multiple languages. Its open-source nature allows researchers and developers to customize the model for specific use cases.

LLaMA has been widely adopted for academic research but also powers Meta’s virtual assistant features on platforms like WhatsApp and Messenger. Its flexibility makes it a favorite among developers looking for cost-effective AI solutions.

In short, LLaMA helps researchers study better with smarter tools.

Code LLaMA by Meta

Noteworthy because of its specific use case, Code Llama is Meta’s specialized model designed for coding tasks such as debugging software or generating scripts based on user prompts. Released alongside LLaMA updates in August 2023, Code Llama supports multiple programming languages including Python and JavaScript.

Developers use Code Llama extensively within Integrated Development Environments (IDEs) for live coding assistance or automated testing frameworks. It simplifies complex programming tasks while improving productivity for software engineers.

In short, Code Llama writes code faster than humans can!

Nova by Amazon

Nova is Amazon Web Services’ (AWS) suite of generative AI models, launched in December 2024. Designed for businesses of all sizes, Nova offers powerful and cost-efficient AI solutions through AWS’s Bedrock platform. The Nova family includes models tailored for text analysis, multimodal tasks, image generation (Nova Canvas), and video creation (Nova Reel). These models are highly scalable, customizable, and affordable—up to 75% cheaper than competitors like OpenAI’s GPT-4o. Nova also supports long-context windows (up to 300,000 tokens) and robust safety features like content moderation and digital watermarking.

Nova is widely used across industries for tasks like creating marketing content, personalizing e-commerce recommendations, automating customer support, and analyzing financial or healthcare data. Its seamless integration into AWS services like SageMaker makes it ideal for enterprise applications requiring scalability and innovation.

In short, Nova helps businesses save money while creating smarter tools to solve large and complex problems.

Claude by Anthropic

Claude was introduced by Anthropic in March 2023 with a focus on ethical AI development through its "Constitutional AI" framework. This model prioritizes safety while maintaining high performance in reasoning tasks. Claude 3.5 Sonnet (released mid-2024) offers enhanced coding capabilities and multimodal features like interpreting photos or diagrams. (It is noteworthy that both Amazon and Google have made multibillion dollar investments in Anthropic, which was founded by former OpenAI employees.)

Claude is especially popular among enterprises for summarizing long documents or generating nuanced responses to complex queries. It is also widely used in industries like healthcare for analyzing patient data and in legal settings for reviewing contracts.

In short, Claude is designed to help businesses answer hard questions safely and clearly.

So what does it all mean?

The short answer is...we don't know yet, but ignore it at your own peril.

At Acuity Advisor Group, we are working hard to help business leaders make smart choices around how to use AI and...just as importantly...where not to rely on it.

Enterprises with a technology component to their offering cannot afford to ignore the benefits of AI in empowering rapid improvements to the customer experience.

Even industries like the trades and services would do well to take advantage of free access to the most transformative technology in our lifetime. It can help leaders better understand large data sets that they would never have time to analyze while attending to the operational considerations of their business.

In this article, we've provided links to each of the technologies. Most of them are free to use at some level.

If you need help understanding how to use them to improve your business, click the button below for a free consultation.