The race for Artificial Intelligence dominance is now on. And with the rise of AI has come dire warnings about its impact on governance and humanity at large. But the challenge that AI represents to democracy is already happening.
The likes of ChatGPT have made recent headlines. But we often forget that our everyday search engines and social media platforms are already enabled by AI and have been for some time. This has already brought mixed consequences for democratic societies.
On the positive side, thanks to AI-enabled social media platforms, spaces like Facebook have expanded the public sphere, where people are better able to participate in the discourse through the unmediated publication and exchange of ideas. With others, such as search engines like Google, we have unprecedented access to information.
However, in democratising our ability to communicate and access information, these AI-enabled platforms also undermined one of the essential pillars of our democracy – journalism. In the space of just a few years the business model, based on advertising, that had been used for countless decades to fund commercial journalism, was upended by AI.
Much of that advertising dollar has now transferred from legacy media to digital platforms. The coalescence of AI and the internet created a market for narrowcast advertising that was far more efficient than its broadcast predecessor. Many existing news platforms struggled to adapt. AI is particularly well suited to the job of deciding which ad to show to whom. The click-through-rate for Google Search ads is 1.91%, so an ad will be clicked on almost one time in 50 that it is shown. The equivalent for newspaper and TV ads is a fraction of this. Companies seeking to advertise naturally enough preferred the efficiency.
Even though this shift has undermined traditional journalism, ironically, access to media has never been more democratic. Everyone can tweet their avocado toast recipes or take to Instagram to stream their karaoke ambitions to the world at no cost. Arguably, the AI-enabled feedback mechanism that determines which missives crowd people’s feeds is far more democratic than the mogul-driven system it replaced. It is hard to argue that net democracy has been decreased as a result.
Even filter bubbles, the bias that skews what information individual users see on the internet, can also be considered as democracy at scale. These bubbles are caused by the algorithms that digital sites use to personalise the user experience (and target advertising). Connecting Internet users with like-minded individuals who share their interests, whether for cat videos or storming the US Capitol, is just an online extension of the book clubs and sewing circles of old. This is democracy at work.
The problem isn’t the filter bubble, or the communication it enables. The bigger challenge is that analysis has been supplanted with opinion. This is not about people with dangerous opinions – to suppress debate is anti-democratic, even if you dislike the opinion being put, or the medium it is communicated over. No, the issue comes back to funding journalism. There are less people investigating, vetting and analysing information professionally.
For all of the talk about AI becoming sentient, taking jobs, or taking over the world, the real danger is that it is creating new services and markets that disrupt the status quo. This leads to unintended consequences, such as undermining the business model by which we funded journalism.
Enabling democracy requires a set of trained professionals who fearlessly and unwaveringly illuminate the failings of a society. Australia, and every democracy, needs to fund journalism properly. The fact that we previously relied on media moguls and their advertising revenue to fund this activity is no reason to return to that model.
Regulation of disruptive AI technologies is insufficient. Uber showed that AI-enabled companies can disrupt highly regulated industries. The previous Australia government’s attempt to ensure Google and Facebook to pay digital advertising tax had little impact.
This is where industry plays a significant role.
Australia’s opinions, regulations, and aspirations surrounding AI are sadly irrelevant because we are simply not an active participant in industrial development. AI research is strong in universities, but industry has serially failed to engage in the opportunity. The result is that Australia is left only with the downside of AI, such as when it enables online multinationals to replace our national information sharing system. We don’t get the upside, such as when an Australian company uses AI to create an entirely new global market, and we can use the corresponding tax revenue to pay for public goods that foster democracy, such as high-quality journalism.
If Australia is to remain a thriving democracy, the country must actively participate in the new AI-enabled global economy. We are not currently. The unintended consequences to democracy of not participating will extend far beyond journalism. Our allies and partners are pulling ahead.