Photo by mundissima / Shutterstock.com
Meta has unveiled a significant advancement in its AI capabilities, introducing a series of updates aimed at deepening the integration of AI across its platforms. The integration of Llama 3, the company’s latest large language model, into its AI chatbot across Instagram, Facebook, WhatsApp, and Messenger indicates that the company plans to compete with OpenAI and Google.
Meta AI is available in more than a dozen countries outside the United States, including Australia, Canada, Ghana, Jamaica, Malawi, New Zealand, Nigeria, Pakistan, Singapore, South Africa, Uganda, Zambia, and Zimbabwe. However, one of the most promising AI markets, India, has yet to see integration.
The company has also launched a new website, meta.ai, which allows users to use the chatbot. “We want Meta AI to be available when you’re trying to get things done at your computer, too, so we’re rolling out meta.ai (the website) today,” said the company.
Among the new features introduced, Meta has enhanced image generation capabilities, enabling users to create and modify images directly within the chatbot interface. Furthermore, users now have access to web search results, expanding the array of tasks that Meta AI can seamlessly handle.
The updated Meta AI offers users various interactive capabilities, from retrieving information within their Facebook feeds to receiving tailored recommendations and assistance.
Users can ask questions related to Facebook posts. For instance, if users see a picture of the aurora borealis on their feed, they can ask Meta AI for suggestions on when to visit countries in the north to see the event or why the event occurs.
Despite these advancements, challenges about content moderation and AI ‘hallucination’ persist. Instances of AI-generated responses veering into nonsensical or inappropriate content underscore the ongoing necessity for effective moderation and AI training protocols. Due to the misinformation and biases, Google stopped Gemini from responding to election queries.
Meta’s strategic focus on AI mirrors broader industry trends, with competitors like OpenAI also pushing the boundaries of AI-powered services.
As companies integrate artificial intelligence tools into their apps and websites, there is a potential cybersecurity aspect to this.
Only a few weeks ago, it was reported that several large language models (LLMs) are vulnerable to ‘many-shot jailbreaking‘. Also, in March this year, it was reported that most generative AIs are bugged and can leak information to hackers.
In the News: Apple removes WhatsApp and Threads from App Store in China