“Initially, AI-driven platforms enticed us with attractive perks…However, as time passed, we were at the mercy of ever-shifting algorithms…leading to diminished income…”

The scenario described above, initially written about ride-hailing apps like Uber and their drivers, could easily apply to social/search platforms and content creators — or to advertising networks and digital publishers.

On a recent International News Media Association (INMA) study tour, Sandeep Amar said that in India, Google and Meta command 85% of the $3-4 billion digital advertising market. Publishers directly access only 3-5% of that market. From any indirectly served advertisement, ad networks retain about ~50% of the revenue. These figures are similar in many other countries.

It makes little economic sense to relinquish ~50% of revenue to AI-driven platforms whose underlying technology is now commonplace. A trusted authority is needed to invest the capital expenditures to level the playing field against the data and network effect moats of these AI-driven marketplaces.

But is it achievable? Has it been done?

Over the past eight years, India has leveled the playing field against Visa and Mastercard in the payments sector with UPI, significantly catalyzing the digitization of the economy. In August 2023 alone, 10 billion payment transactions were made via UPI. UPI’s widespread use and acceptance are such that even companies like Google, Amazon, and Uber have adopted the open protocol.

India is now attempting to replicate this success in e-commerce and mobility with ONDC and Beckn, respectively. In these endeavors, they have not limited Visa, Mastercard, Amazon, or Uber in any way — just created an open alternative of comparable size. This allows new startups and non-incumbents to compete.

This approach has greatly benefited merchants and cab drivers. For instance, unlike Visa and Mastercard, UPI charges no transaction fees from merchants or consumers. In Bengaluru, Namma Yatri, using the Beckn protocol, charges a 2% commission, compared to the 20% charged by ride-sharing apps.

Why adopt this solution for the news industry?

Digital news firms are likely to experience a significant drop in revenue due to third-party deprecation, fluctuations in search and social algorithms adjusting to AI-generated content, and audiences shifting from search to chat interfaces. News firms have limited means to combat this.

The truth is AI-driven marketplaces operate vastly more efficiently and relevantly than current news firm operations. For example, a Profitero study shows that Amazon changes prices on its site 2.5 million times a day. In comparison, editorial teams update story listings on their websites and apps far less frequently.

We need to build ourselves out of this situation.

And yes, we should invest in news long-term! The demand for news — verified and vetted information — will always exist. Much like payments, commerce, and mobility, news is a fundamental civic infrastructure. Moreover, advertising is an economic reality. Both will persist. The question is whether news and advertising on news will survive with news firms in their current form.

I predict that governments and the broader media development ecosystem will recognize the value in creating open systems to build viable alternatives to BigTech’s advertising, search, and social platforms.

Open systems for…?

We are likely to see two open systems.

Open DMP: The first will be an open protocol allowing collectives of willing publishers to pool first-party data into a common Data Management Platform (DMP), and then utilize off-the-shelf or custom targeting algorithms to sell programmatic advertisements. To seed demand, governments could mandate that a certain percentage of their annual advertising spending be served via this DMP. By default, the open DMP would assign a publisher to serve an advertiser’s campaign needs but would also offer options to select from other publishers. This would allow publishers to serve their direct sales campaigns via the Open DMP platform.

Open News Verification as feedback to the Open DMP: Offline, brands pay a premium for their advertisements in well-researched and reputable newspapers and TV shows. However, in ad network advertising, there’s no differentiation between a funny cat video and a well-researched piece of journalism. An impression is just that — an impression! An open platform for verification and vetting built on fact checks from Poynter’s IFCN or India’s Misinformation Combat Alliance, along with datasets or initiatives like The New York Times’ Provenance project, could be used to do pre-publishing checks, like YouTube’s Content ID copyright check, and post-publishing fact-checks, like Twitter’s Community Notes. Publishers that contribute to and mandate the pre-check and showcase the Community Notes on their platform could demand a higher rate for their content as compared to advertisements shown next to an anonymous blog.

What’s remains unknown

Payments, commerce, and mobility are civic infrastructural systems that impact the daily life of hundreds of millions of citizens, mom-and-pop stores, and drivers. It makes practical sense for the government to invest in systems for them, which will have a direct positive ripple on the economy. Would the government or civic funders see tangible value from the effort and money poured into leveling the playing field for news?

Today, the algorithms in AI-driven marketplaces optimize revenue by serving advertiser needs while barely meeting supplier (publisher, cab driver) needs. If today your earnings are $10, over time the algorithm will automatically test if $9.90 is acceptable. At a cumulative level, humans might not perceive this drop as significant, but like inflation, a gradual decline in rates will add up over the years.

In contrast, the open DMP would need to adopt a participatory approach and balance the interests of all stakeholders — news publishers and brands. These are challenging problems with no easy solutions, but they’re worth tackling for the information infrastructure of our civic systems.