WOULD you trust artificial intelligence (AI) to help run your country? Malaysia is doing just that, turning to AI to help with some of the more tedious parts of governing.
The South-East Asian nation announced a plan to roll out Google’s Gemini AI tools to nearly half a million civil servants as part of a broader top-down push to help make its workforce more efficient.
The stakes are higher when it comes to relying on this technology to help run a democracy where people never elected machines into office.
There are still many examples of AI messing up or going rogue. But the reality is that AI is likely coming for swaths of repetitive office work, including within government. Malaysia is not alone in seeking ways to automate some of these tasks.
Yet, it still seems like quite a bit could go wrong. Within days of Kuala Lumpur’s announcement, Google launched a Super Bowl ad that touted how Gemini could help a cheese-shop owner in the United States grow their business online.
In the commercial, however, its AI programme spit out an inaccurate statistic about Gouda consumption. And this isn’t the first time Google’s generative AI products have run into trouble.
If these tools are getting it wrong when it comes to cheese, it’s hard to imagine they are ready to draft policies.
Despite these concerns, Malaysia seems to be doing a lot of things right. It’s going slowly, working with stakeholders to build trust, and offering transparency to the public along the way.
The government launched a National AI Office to work with the private sector to implement the new technology. The office has laid out near-term goals and committed to publishing report cards on its progress.
Before bringing Google’s AI to 445,000 workers, Malaysia also launched a smaller test run. And officials are working on public sector AI adaptation guidelines, expected to be released later this month.
The government said it hopes that workers will be able to use these advanced tools to help draft policy papers, analyse data, enhance public engagement and save time.
It’s a far cry from what some prominent researchers are calling an “AI coup” happening in Washington under Elon Musk’s Department of Government Efficiency or DOGE (whose acronym is a meme cryptocurrency).
Despite sowing widespread chaos across the federal government under the banner of cutting waste, little is known about the people running DOGE who have been given this immense power to upend agencies.
The group has a website tallying supposed savings, but even this has faced criticism for errors and not offering nearly enough clarity about its processes.
There seems to be an over-arching AI strategy at play, with a goal of replacing as many human jobs as possible with machines.
But none of this has ever been clearly laid out. And if that is the case, Musk and the Trump administration must take this process much slower in order to build and maintain trust.
The billionaire has taken to posting updates about DOGE’s work on X, the social media platform he owns, and he is also spearheading an AI startup.
Delegating swaths of government work to AI also gives new power to the people in the tech industry creating and controlling these machines.
In Malaysia, we at least know that Google’s products are being used, and local media was invited to demonstrations about how they can be employed.
The products and intentions of Musk’s overhaul, meanwhile, remain cloaked.
Some scientists claim they’ve been warned that AI is being used to detect “forbidden words” in federal research funding applications, such as “diversity, equity and inclusion” and “climate change.”
Others suggest Doge is working to rapidly develop a custom AI chatbot for use in the US General Services Administration. If so, people deserve to know how this is being programmed.
Ultimately, AI still isn’t up to the task of governing. Writing new policies, defending them to the public, and implementing them still takes an immense amount of critical human decision-making. It’s hard to imagine America’s founding fathers using a chatbot to write the Declaration of Independence or the Constitution.
There are still ways that AI tools can be used to assist in the workplace and boost productivity. But right now much more attention will be placed on the emerging technology’s missteps – especially when it comes to government use.
Thus any nation’s rollout across federal agencies will inevitably be closely scrutinised and could make or break the tech’s future progress even in areas where it could be beneficial.
At least for now, Malaysia’s approach is aimed at providing civil servants with “gen AI assistants,” and not replacing the workers. The goal isn’t to remove the humans, but to free them up to focus on more complex tasks. This approach has a greater chance of success, though authorities must maintain transparency along the way.
As more countries inevitably turn to AI to help with federal efficiency, they would be wise to remember that to move fast and break things might work in Silicon Valley, but not in governing. — Bloomberg
Catherine Thorbecke is a Bloomberg Opinion columnist. The views expressed here are the writer’s own.
Source: The Star