Meet the A.I. Jane Austen: Meta Weaves A.I. Throughout Its Apps


Meet the A.I. Jane Austen: Meta has integrated artificially intelligent characters, such as Jane Austen, Snoop Dogg, and others, into Instagram, Facebook, and WhatsApp as part of the intensifying competition to lead the technology race.


In a recent WhatsApp text conversation, we engaged in a dialogue with Jane Austen, the renowned 19th-century British author, inquiring about her thoughts on Mr. Darcy, a character from her celebrated work, "Pride and Prejudice."


After a brief pause, Ms. Austen responded.

"Ah, Mr. Darcy. He is remembered by everyone as one of my characters," she expressed, her face displayed in a small window above our conversation. "But fewer individuals have actually read one of my books," she added, raising her eyebrow with a slight touch of resentment.


However, it was not Ms. Austen herself conversing with us. Instead, Meta, the parent company of WhatsApp, Facebook, and Instagram, employed a modern interpretation of her likeness as part of an artificially intelligent character capable of engaging in conversations across the company's messaging applications. Similar character representations based on various personalities, such as former quarterback Tom Brady, social media influencers Mr. Beast and Charlie D'Amelio, and hip-hop artist Snoop Dogg, were also available for interactions.


These characters are part of a suite of products recently introduced by Meta. These products, powered by artificial intelligence, are set to be extensively integrated into platforms like Instagram, Messenger, and virtual- and augmented-reality devices such as the Quest 3 headset and Ray-Ban Stories smart glasses. Additionally, the release includes a chatbot that will leverage Microsoft's Bing search engine and A.I.-assisted image-editing tools designed explicitly for Instagram usage.


Mark Zuckerberg, Meta's CEO, emphasized the limited exposure most people have had to the latest and most potent A.I. technologies during the announcement. "That's something we aim to change," he stated. Furthermore, he highlighted the preference of users to interact with multiple intelligent A.I. entities rather than a single, super-intelligent one.


Meta is striving to keep pace with OpenAI, Google, Microsoft, and other prominent companies vying in the rapid race for A.I. capable of instantly generating text, images, and other types of media. Since the unexpected launch of OpenAI's chatbot ChatGPT in November, Silicon Valley executives have recognized the technology as the next significant computing paradigm shift and have worked tirelessly to remain competitive.


For Meta, widespread acceptance of its new A.I. products could considerably enhance user engagement across its numerous applications, a majority of which depend on advertising for revenue generation. Increased time spent within Meta's apps would correspond to an increased number of ads presented to its user base.


Although Meta has invested years in A.I. research behind the scenes, it initially lagged behind in introducing generative A.I.-based products. This delay was partly due to Meta's focus on transitioning into a metaverse company. To catch up, Mark Zuckerberg implemented substantial changes within the organization, redirecting its efforts toward A.I.-centric products like those unveiled on Wednesday. Weekly meetings with Meta's executive team were conducted to evaluate progress and steer the company's transformation.


Now, Meta is leveraging its colossal global user base, surpassing three billion individuals, to facilitate the mass adoption of its offerings. While prior interactions with ChatGPT and Google's Bard chatbot mainly occurred on an individual level or within productivity applications like Gmail, Mr. Zuckerberg envisions users interacting socially with Meta's chatbots across Instagram, Facebook, and WhatsApp in everyday conversations and group chats with friends.


For instance, suppose two friends utilizing WhatsApp wish to discover a suitable place for a Sunday morning breakfast in San Francisco. In that case, they might seamlessly turn to Meta's newly introduced chatbot, Meta AI, and inquire, "Where can we find the best place for eggs Benedict in San Francisco?"


Within seconds, Meta's digital assistant would tap into Microsoft's Bing search engine for real-time web results, leveraging its underlying A.I. capabilities to deliver a concise and conversational response. Individuals seeking more detailed information could access linked footnotes that direct them to Bing search results in a separate window.


Other social networking applications have also begun integrating A.I. functionalities. Snapchat, a popular platform among teenagers, integrated a chatbot this year, manifesting as a customizable digital avatar. Additionally, Character.AI, a Silicon Valley startup, provides a service allowing individuals to engage in conversations with a realistic facsimile of almost anyone, whether living or deceased.


Meta's innovative character-like bots operate in a similar manner to Snapchat's offering and offer users a glimpse into the digital personalities offered by Character.AI.


One Meta A.I. character, named Amber and modeled after Paris Hilton, assumes the role of a detective partner in chat interactions, specializing in solving whodunits. Joining the lineup is Kendall Jenner's character, Billie, described as a loyal and straightforward companion who doesn't tolerate nonsense. Meanwhile, tennis star Naomi Osaka's character, Tamika, embraces her love for anime as an aspiring Sailor Senshi. Snoop Dogg takes on the role of a Dungeon Master in a playful homage to Dungeons and Dragons, offering an interactive experience where users can choose their own adventure.


These celebrity and athlete A.I. characters receive compensation, although specific details have not been disclosed by Meta. The characters provide a means for people to have fun, acquire knowledge, share recipes, or simply pass the time, all while fostering connections with friends and family, according to company executives.


Ahmad Al-Dahle, Meta's vice president of generative A.I., explains that diverse A.I. offerings cater to different needs, laying the foundation for an expanding ecosystem of various A.I. options in the future.


However, similar to other chatbots, Meta's A.I. chatbots have the potential to generate false or misleading information, a phenomenon known as hallucination among researchers. Even when the bot bases its responses on information retrieved from search engines, inaccuracies can occur.


As an example, when asked about the winner of the 1904 World Series, the Meta AI assistant mistakenly answered "the New York Giants" before offering further assistance. Subsequently, when inquired about the existence of the World Series in 1904, the bot corrected itself, acknowledging that the Giants refused to participate, resulting in no Series that year.


In addition to its chatbot capabilities, Meta, like Microsoft, OpenAI, and other companies, provides a tool for instantly creating photorealistic images. Users can quickly generate A.I.-produced photos or sticker emoji reactions within the company's messaging apps by inputting text into the chat prompt. Some limitations apply to this image-generating technology, which, if misused, could contribute to the spread of disinformation online. To combat this, Meta ensures that images produced using the tool are clearly labeled as A.I.-generated.


Mr. Al-Dahle emphasizes Meta's commitment to responsible use of their technologies. Thousands of hours have been dedicated to conducting "red team" scenarios to identify and mitigate potential misuse. The company has also published guidelines for responsible use, inspiring individuals interested in leveraging Meta's underlying technology to create their own chatbots.


During the initial phase, many of Meta's products will be introduced exclusively to U.S. users. This allows the company to address any early issues and observe user feedback.


The widespread presence of chatbots may become a reality sooner than anticipated. Meta's recent public release of LLaMA 2, the underlying code for their advanced A.I. technology, saw an enthusiastic response from developers, with over 30 million downloads to date. Collaborating with Microsoft, Google, and Amazon's cloud services divisions, Meta intends to host the technology, enabling developers to build the next generation of versatile bots capable of fulfilling their specific requirements, both beneficial and potentially problematic.


As for the A.I.-powered Ms. Austen, she shares her wisdom selectively. When asked about the ideal age for a woman to marry, she tersely declines to provide an answer. Instead, she humorously remarks, "Do you expect me to dictate your love life? Marry whenever you find someone who can embrace your eccentricities as you embrace theirs."