Image: cokada/Getty Images.
Meta announced that its generative artificial intelligence (AI) advertising tools could not be used to fuel political campaigns anywhere in the world, with access blocked for ads targeting specific services and issues.
The social media giant said earlier this month that advertisers would not be able to use its ad manager’s generative AI tools to produce ads about politics, elections, housing, jobs, credit or issues. of a social order. Advertisements relating to health, pharmaceuticals and financial services are also affected.
This policy will apply globally as Meta continues to test its generative AI ad creation tools, confirmed Dan Neary, Meta’s vice president of Asia Pacific. “This approach will allow us to better understand the potential risks and put in place adequate protection measures for the use of generative AI in advertising related to potentially sensitive topics in regulated sectors,” he specifies. .
A launching pad for serving ads
The move comes as several important elections take place next year – including presidential elections in the United States, Finland, Pakistan and Taiwan.
Meta – whose social media platforms include Facebook, Instagram, WhatsApp and Threads – has made AI a top priority and plans to add generative AI capabilities across all of these platforms.
Its Ads Manager is presented as a launching pad for serving ads across platforms. It offers an “all-in-one tool” for creating, managing and tracking ads. A sandbox was introduced in May to allow testing of Meta’s new generative AI tools for advertisers. Among them are background generation, text variations and image cropping. With text variations, for example, advertisers can generate multiple versions of text to reach different audiences.
More than half of advertisers use AI tools
The company also unveiled an AI-powered chatbot, Meta AI. It includes an AI image generation tool called Emu. These images can then be used on WhatsApp and Instagram.
Asked about the adoption rate of its AI products, Dan Neary explains that more than half of advertisers use the company’s Advantage+ tools to optimize the images and copy of their ad creatives. These tools have helped advertisers achieve $10 billion in revenue from Advantage+ shopping campaigns, he added, citing figures reported by Mark Zuckerberg, CEO and founder of Meta, at the recent publication of the company’s results. The number of advertisers using Advantage+ shopping campaigns each week has tripled compared to six months ago.
Further emphasizing the role of AI, the vice president noted that 20% of Facebook and Instagram news feed content is now recommended by AI.
Show more relevant content
There was a concerted effort over a year ago to display more relevant content powered by recommendation engines, rather than content organized around people followed by Meta users. AI also helps drive better results for marketers, with tools like the Advantage+ suite automating their tasks, Neary says.
“We believe that every connection is an opportunity for businesses and we see this across all our platforms,” he says. Some 3.96 billion people use at least one of Meta’s services every month, 3.14 billion of them daily. Around 40% of users reside in Asia Pacific, where mobile consumption is high, especially for messaging services.
Asked about how data from its products and services is integrated and used to train its generative AI tools, Dan Neary explains that a variety of sources are used: “Generative AI models require a large amount of data to be trained effectively, which is why a combination of sources is used for training, including information publicly available online, licensed data, and information from Meta’s products and services. »
” To do things well “
When it comes to publicly available information online, datasets are filtered to exclude “certain websites that typically share personal information,” Neary points out. Posts shared publicly on Instagram and Facebook, including photos and text, are part of the data used to train the generative AI models that power the features announced earlier this year.
“We didn’t train these models using people’s private messages. We also do not use the content of private messages you send to friends and family to train our AI tools,” he reassures. “We may, however, use data from your use of AI-created stickers, such as searching for a sticker to use in a chat, to improve our AI sticker models. »
As for how Meta is responding to growing concerns about AI security and the use of personal data, Dan Neary says a specialized interdisciplinary team is responsible for ensuring its technology is designed and used in a timely manner. responsible manner. The team also gathers feedback from external experts and regulators, he says.
With generative AI still in its infancy, he said Meta was now working to collaborate with key industry stakeholders to “get it right”.
Source: ZDNet.com