Can Meta's AI Assistant Be The Game Changer in An Already Saturated Chatbot Market?

A couple of weeks ago, Meta launched its AI assistant tool, Meta AI, across all its platforms (i.e., Facebook, Messenger, WhatsApp, and Instagram) in its biggest push to make AI a core feature of all its apps, This can be seen as Meta’s response to OpenAI’s ChatGPT, Google’s Gemini, and Microsoft’s Bing AI. Apart from the U.S., the AI assistant tool has been unveiled in more than a dozen countries including Australia, Canada, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, Zambia, and Zimbabwe. 

But in an AI chatbot market that’s already heavily flooded by many AI assistant tools, can Meta AI prove to be a game changer or at least make its presence felt? Well, you are about to find that out in this blog.

So, without wasting much time, let’s take a look at what Meta AI has to offer and whether or not it can be the next big thing in the realm of AI assistant tools.

What is Meta AI?

Meta AI is a free tool baked into apps that millions use daily. Its latest version leverages Llama 3, Meta’s latest large language model. While it is quite similar to Open AI’s ChatGPT Plus with regard to user experience, it comes without the latter’s $20-per-month price tag.

For example, Meta AI’s built-in image generator, Imagine, can dish out pictures based on text prompts, similar to ChatGPT Plus. While it is no match to the superior graphics of ChatGPT’s DALL-E 3, it can churn out images in real time, letting you see the elements fall into place one piece at a time as you type the prompts. It can even figure out prompts mid-word and give you what you want even before you ask for it.

However, since the technology is new, it might not always return accurate responses, which is the case with other AI systems too. Let’s take a look at what works and doesn’t work in Meta AI.

Factual Accuracy

When asked about the ongoing conflict between Israel and Iran, Meta AI came up with a concise and bulleted list that included dates. But it cited only a single CNN article. To crosscheck if Meta AI was enlisting the help of Copilot’s AI model, which Microsoft borrows from OpenAI), I typed the same query on Copilot, which also came up with a bulleted list with roughly the same info but more citations.

Meta AI’s response was factually correct and up-to-date, although not particularly eloquent. Also, the mobile response was much more compressed, which means that you could be getting a truncated answer.

Copyrighting Capabilities

As one would expect after using ChatGPT and Gemini, Meta AI can dish out page after page of content from its content stockpile. After all, Meta has an endless supply of captions to train on. So, when I asked it to generate marketing copy suggestions for an imaginary clothing brand, it churned out multiple suggestions, which only goes on to show how rote this task has become.

One can only feel sorry for marketers! Their hard work penning millions of pitches for their brands has defined a style that AI has perfected over the years.

Editing Skills

Meta AI excels at editing existing paragraphs. For example, when I asked it to edit paragraphs that felt verbose, it trimmed all the unnecessary words. When I requested it to improvise sentences written in passive voice, it rewrote them in active voice and also added more context.

I wanted to see if it could simplify high-jargon content and to my surprise, it removed jargon from a tech blog that I provided and rewrote highly technical terms in plain language.

It’s Not a Search Engine

Meta said its chatbot could be a replacement for web search. Honestly, I would advise you against seeing this AI assistant tool as anything more than a chatbot since it fails spectacularly at basic search queries like looking up airfares and weekend activities.

When I enquired about flights from New York to Colorado, Meta AI generated output on how to take public transportation from the Denver airport to downtown. Upon enquiring about flights from Oakland to Puerto Vallarta in Mexico, the bot listed flights departing from Seattle, San Francisco, and Los Angeles.

Also, it makes up other basic information. For instance, when I enquired about where one can go for a romantic weekend in Oakland, its response included a fictional business. Gemini and Copilot did much better at these types of search tasks. However, nothing beats clicking on a link through an old-fashioned web search.

Meta AI’s Integration With WhatsApp

Meta AI’s integration with WhatsApp allows people to leverage AI to access real-time information, such as news updates or weather predictions. It enables people to seamlessly go between AI tasks and conversations, thereby, facilitating a more intuitive user experience with generative AI.

However, privacy and security remain a concern, especially considering WhatsApp’s commitment to end-to-end encryption. This means that customers might not feel too comfortable with providing private or sensitive material to an AI algorithm, which could result in data security issues and user authorization.

Bottom Line

Honestly, I am not a big fan of AI assistant tools owing to my old-school thoughts on scouting through multiple online sources to deliver content. But I understand how useful AI chatbots like ChatGPT and Gemini can be for writers who don’t want to do the hard grind. However, I didn’t find Meta AI particularly useful for anything. In fact, Microsoft’s Copilot works much better as a free AI assistant tool than Meta AI.

However, in case you are too lazy to open the browser to look for “best Chinese New Year captions” or “witty wedding wishes”, you can enlist the help of Meta AI. TikTok won’t help you with that. Yet!