Regulating AI

SINGAPORE ­– Filipinos aren’t the only ones worried about the use of AI-generated deepfakes to manipulate elections. And we’re not the only ones concerned about possible Chinese interference in our elections through disinformation.

Other countries are doing something about it. Last year, for example, South Korea revised its Public Official Election Act and banned deepfake videos, photos and audio for election campaigns within 90 days before election day. But Chat GPT can still be used to write campaign speeches and song lyrics or generate an election slogan. Violators face imprisonment of up to seven years or fines of up to the equivalent of $37,500.

Private South Korean AI chatbot service providers also stepped in, jointly committing to deter the malicious use of deepfakes for elections. One launched a system to create watermarks on AI-generated content. Another vowed that its AI chatbot service would not generate inappropriate content and promised close monitoring of its platform. A generative AI firm worked with the National Police Agency to customize a system of ferreting out deepfakes and improving response to election-related offenses.

South Korea has confronted not only North Korean but also Chinese disinformation and propaganda operations during elections in 2020, 2021 and 2023 that aimed to promote pro-China and anti-US sentiments.

Singapore, which has taken note of the AI-related measures undertaken by South Korea, is also moving to prevent the use of generative AI to undermine elections.

Josephine Teo, Singapore’s minister for digital development and information, told a gathering of the Asian Journalism Fellowship (AJF) here last Friday that Singapore has passed a law initially tackling fake news and passing opinion as fact. Content is not removed, but a label is pinned on it to inform the public.

In our country, even such labeling or tagging, however, can be controversial and may raise concerns about free expression.

Teo said among the tasks of her ministry are to promote digital inclusion, make rules interoperable and promote “AI for the public good for Singapore and the world.” The ministry is of course monitoring the impact of AI on jobs, especially on “knowledge workers” such as professional managers.

The aim is to make AI help people work more efficiently, she said.

“What do we do with (AI)?” Teo asked. “Deal with it. It’s here to stay.”

*      *      *

Singapore has always been open to new technology. But for many people, the simplest stuff can be intimidating.

In this high-tech city-state, there’s no more telephone unit in my hotel room. Instead a device on the desk like a cube-shaped alarm clock is used for phone calls, and has an AI-enabled voice assistant.

To “wake up” the voice assistant, you tap the device and say “Hello AI” or its equivalent in Mandarin, followed by your command, such as “call front desk” or “what time is check-out?”

I followed the step-by-step instructions to set the alarm, but the darn device wouldn’t obey my voice command; maybe it doesn’t recognize Pinoy English. So I ended up asking a human to give me my wake-up call.

Technology can make you feel like an idiot. But feeling stupid is still better than feeling that you’re headed for extinction.

I’m in this city-state for one of an ever-increasing number of international media gatherings tackling the impact of generative artificial intelligence on journalism.

AI was the main topic at the 2024 International Media Conference of Hawaii-based East-West Center, which was held in Manila last June with the theme, “The Future of Facts.” I recently skipped another AI-focused media seminar overseas. It was the topic at the AJF editors’ forum that I attended here from Friday until yesterday morning, sponsored by Singapore’s Temasek Foundation and the Institute of Policy Studies at the National University of Singapore.

Teo, along with the speakers and panelists at the gathering, were one in saying generative AI is not about to make journalism extinct, despite what doomsayers have described as a global “media apocalypse.”

But generative AI – or artificial intelligence that can create content such as music, images and stories – has made the journalist’s job of sifting fact from fiction more challenging, particularly come election season.

*      *      *

With the easy spread of disinformation, “there is an increasing need for verified, independent information,” according to professor Charlie Beckett, director of Polis, the journalism think tank of the London School of Economics and Political Science.

A global survey conducted by Polis last year showed that about two-thirds of journalists worldwide are already actively using some form of generative AI – but only for support in their work.

Even as governments consider rules to prevent the malicious use of AI, both Beckett and Jeremy Au Yong, associate editor for newsroom strategies at Singapore’s Straits Times warn about regulating AI.

“Be careful what you wish for,” Beckett said, as he noted that among the first to support regulation of social media are authoritarian rulers.

Chua Chin Hon of Singapore’s national Mediacorp News Group, which has been using AI in the newsroom for over three years now, warns journalists to “be careful breaking what you cannot fix.”

To illustrate, he noted that for a long time, media organizations worked on the idea that online news is supposed to be free.

“Now we’re trying everything to unbreak that,” he said. “If news is free, good luck trying to get another subscription.”

Yong is hoping that traditional media outlets can negotiate as a bloc with tech giants such as Google so that proper compensation can go to the generators of content.

At the same time, about 12,000 journalists worldwide are part of a network pursuing a global initiative that will empower news organizations to use AI responsibly.

Teo, addressing the media gathering last Friday, said that with the spread of deepfakes and disinformation, “all the more we need quality journalism; all the more we need the people who help us make sense of the world.”

The trick in making generative AI a positive tool, according to Chua Chin Hon, is to “move fast without breaking things.”