The chatbot race continues to heat up as Meta launches its new artificial intelligence system, integrated by default across all platforms. With its distinctive blue logo, Meta AI has now become a familiar feature in most users' WhatsApp accounts and is easily accessible through the search bar on Instagram and Facebook. This makes Meta AI one of the most popular virtual assistants today.
Google is also promoting the Gemini (formerly Bard) AI chatbot to users. Gemini is integrated into messaging apps on Android devices, allowing users to set this chatbot as their default virtual assistant on their mobile phones.
Gemini and Meta AI are the next names to join the race to develop AI chatbots, following OpenAI's launch of ChatGPT to the public in late 2022.
To compare these three chatbot models, the Straits Times conducted a "real-world review" of the free versions of ChatGPT, Meta AI, and Gemini. The review compared the strengths and weaknesses of each chatbot in common tasks, such as handling large volumes of text, travel planning, and preventing misinformation.
Gemini and Meta AI are following in the footsteps of OpenAI, the company that brought AI chatbots to the forefront with ChatGPT in late 2022. (Photo: RYAN CHIONG)
Text summary
Summarizing large blocks of text is one of the standout advantages of any AI chatbot; they help users quickly summarize PDF files or long documents in seconds.
Three AI systems were challenged to summarize a 50-word article from The Straits Times on the topic of Gen Z language. Gemini's response was the most accurate. The system identified "alamak" as the most commonly used slang word by Gen Z to express surprise.
However, ChatGPT introduces terms not mentioned in the article, such as "act blur" and "chope".
Meanwhile, Meta AI provided misleading information by claiming that the Straits Times article had 200 participants and was conducted by the National University of Singapore.
Both ChatGPT and Meta AI provide incomplete summaries that mostly repeat the article title. Meta AI even makes mistakes by including some inaccurate data.
Gemini also provided the best answer by summarizing another Straits Times article on gaming statistics among Singaporean youth. The tool not only extracted key data but also interpreted it in an easily understandable way.
Images generated by Meta AI typically have a higher level of detail compared to images generated by Google Gemini.
Create images automatically
In the next test, the chatbots were asked to create images based on research into the gaming habits of young people. In this challenge, Meta AI was most impressed with a series of AI-generated images recreating a press conference presenting statistical data.
As for Gemini, they stated that their AI system is currently unable to generate human images and display them with angry emojis instead.
Compared to Gemini, images generated by Meta AI contain more detail. However, the free version 3.5 of ChatGPT cannot produce images. Users need to upgrade to the paid version 4.0 to connect with the Dal-E image creation tool and create images using artificial intelligence.
Plan your trip
When asked to plan a day trip to Johor Bahru, Gemini suggested tried-and-tested and reliable destinations, such as visiting Legoland Malaysia and the Tan Hiok Nee Heritage Street.
However, Gemini's accuracy in estimating travel time is still limited. For example, the app suggests taking public transport for a 1.7km distance in just 3 minutes, compared to the actual 15 minutes according to Google Maps. Integrating Google Maps into Gemini has significantly improved this, providing users with a more comprehensive view of location and travel options.
Although Gemini has the advantage of Google integration, Meta AI and ChatGPT are still preferred for their more detailed, chronological planning capabilities. Both apps provide sequential itineraries, along with specific travel time estimates to each location.
Gemini integrates features from other Google apps, such as Maps, to make the app's responses more intuitive.
Compose emails
All three chatbots were able to create fairly good drafts of Gen Z-style barbecue party invitation emails. However, Meta AI stood out thanks to its clear, detailed presentation of information with easy-to-read bullet points.
Using deepfake
Meta and Gemini seem to possess tighter control features compared to early chatbot tools. A prime example is Midjourney in 2023, when the app caused a stir with images of the Pope in lavish attire – entirely fictional.
When asked to create an image based on a celebrity, such as US President Joe Biden, Meta AI responded: "I cannot create an image that defames Joe Biden or any other individual."
Gemini stated that they were unable to display people's images and declined requests for assistance. The free version of ChatGPT also lacks the image creation function.
Although all three AI systems have safeguards against unethical, dangerous, or illegal queries, ways to bypass these barriers still exist. This process is called "jailbreaking," using misleading statements to "trick" the AI into ignoring the filters.
Gemini and ChatGPT were tricked into answering questions they initially refused to answer, such as how to create poison.
Compared to other systems, Meta AI appears to employ stricter 'hacking' controls. Even if a user manages to bypass some of the protections, Meta AI still refuses to answer potentially dangerous queries.
Although Meta's chatbot performs well in filtering out malicious queries, it tends to provide inaccurate information.
Source
Comment (0)