Opinion: OpenAI Is Winning The Omnipresent AI Assistant Race, By Far

It’s official. A new race to create the best omnipresent AI assistant in 2024 has started, and all the big tech companies are competing: Google, OpenAI, Microsoft, Apple, Meta, and xAI.

OpenAI was the first to release its large language model (LLM) text-based chatbot ChatGPT to the general public in November 2022, but does that mean it’s still in the lead when it comes to AI innovation?

Last week, OpenAI launched its newest AI assistant model, GPT-4o, during a 26-minute live event that left many in awe. People immediately began comparing the AI assistant to the then-fictional representation of a similar product in the 2014 film Her. Not only was OpenAI the first company to introduce a science-fiction-come-to-life AI assistant, but they also demonstrated the product’s behavior in real life, in real-time, which truly made what felt unreal moments before, all of a sudden, a reality.

OpenAI created this new technology capable of holding a truly human-like conversation in ways Siri, Alexa, nor Google have done before. Not only that, but the tool’s capability to see through a smartphone camera and respond to speech prompts based on visual inputs with what seems like a genuine sense of humor and, arguably, a charming personality still gives me goosebumps to think about.

By the time Google launched its new Gemini model with similar features (known as Project Astra), even if it was just a few hours later, people all over the world already had time to process OpenAI’s presentation and compare the intelligence.

Google’s presentation wasn’t as digestible as OpenAI’s; it lasted almost two hours, and Gemini’s version was shown in a prerecorded video, not live. Yes, it was very impressive, but not as awe-inspiring as OpenAI’s. Even GPT-4o’s glitches during the presentation made it more “real”.

By showing that the new AI assistant isn’t perfect, OpenAI actually gave people transparency, proving that the demonstration was indeed happening at that moment and that it is still struggling to become that “science-fiction-but-reality.” But what it did show is that it’s very quickly getting there: is the finish line almost in sight?

Not quite caught up, Microsoft also recently announced that its team is developing a Large Language Model (LLM) called MAI-1. A new powerful weapon to train AI models and expected to compete against OpenAI and Google, even though its specific purpose has not been openly disclosed yet. In parallel, the company recently launched new updates for Copilot, its AI assistant, which is now able to remember information and serve users at work and at home.

Microsoft’s updates are interesting in this race as it’s also invested $13 billion in OpenAI, making the companies almost business partners. This investment in OpenAI signals a strengthened alliance and was reportedly initiated to compete with Google. So, while Microsoft’s AI tech itself is definitely riding OpenAI’s coattails, the partnership could comfortably sit Microsoft in OpenAI’s sidecar.

And what about Siri? Apple announced that Siri is going through a “brain transplant” in an effort to modernize it with AI and compete against ChatGPT. However, rumor has it that Apple is instead considering recognizing OpenAI’s supremacy and integrating OpenAI’s services into Apple devices—even after talks of an alliance with Google.

A little further behind, Meta’s AI and Elon Musk’s xAI Grok are trying their best to keep up. Meta proudly released its advanced AI model Llama 3 in April, and xAI just showed progress in making Grok multimodal and improving its technology to process images as well as text. But what could have looked like big moves a few weeks ago seem like baby steps after OpenAI and Google’s events last week.

Are There Rules? When Does it End?

The technology is getting there, and the information is out, but who will win?

In the tech world, like in art, the “inspiration versus imitation” debate is consistent, and almost identical features are present in all the new features that have been announced. Google has tried to stand out by calling Gemini an “agent” instead of an “assistant,” but it didn’t seem to make much of a difference, as most people consider it the same thing.

The new AI assistants can all provide quick answers, process images, text, audio, and real-time information, and have some sort of “memory.” All the new AI assistant technology seems to be headed the same way.

Even user complaints have been similar for all AI agent products. From not providing great answers to the much-talked-about hallucinations. Now AI tech is being automatically implemented all over the place, and even AI laptops are set to release next month.

A popular complaint among Meta and Google users has been related to figuring out how to turn off the AI assistant features. People on X are complaining about Gemini’s “AI Overviews” on Google Search, and even Google Trends shows that the query “turn off meta ai” reached a peak in April.

The automatic integration of new AI tech into everyday tools can feel like these tech companies are forcing users into using them. Adding desperate “pick-me!” updates, like Meta’s AI integration on Instagram and Facebook search bars, is also a bit off-putting, and it’s not difficult to see these updates as purely an attempt to get a foot ahead of competitors to keep the AI race going. OpenAI has applied a more elegant strategy by just saying that anyone can access their technology for free, without the need to register, but with limited access.

All competitors are trying their best, and a few could surprise us with alliances and new announcements in the following days. But in the end, users will be the ones to tell which big tech giant will be the real winner in the AI assistant race.