Thinking About AI and the Business of News

Four questions to be asking now

Welcome to Second Rough Draft, a newsletter about journalism in our time, how it (often its business) is evolving, and the challenges it faces.

A bit more than a decade ago, I wrote a very short ebook called Why American Newspapers Gave Away the Future. It’s about the crucial moment, centered in 1995-96, when the initial business model for internet publishing was set, with the decision to offer nearly all content for free, and the tragic consequences for journalism that I believe ensued.

I think the little book holds up pretty well, which is nice, but that’s not what this week’s column is about. Instead, it’s about the analogous moment I think we are in now, as publishers grapple with the implications for their business of generative AI. The biggest question is how to make sure not to give away now what we salvaged from the mistakes of the past.

Can we learn from our mistakes last time?

At this early date, I think it’s easier to spot the questions we need to be asking than to be confident of the answers, so that’s how I’ve organized my thoughts:

Do we have the right talent?

One of the root problems in the 1990s was that publishing companies literally didn’t understand what was happening around them. Too few of their editors and executives had any familiarity with the emerging technologies, too few took time to find out, too little effort was made to push back on the constraints binding those who saw the need for change. (Contrast these two takes on one of the outstanding exceptions.)

The few large and well-resourced news organizations that remain won’t make this mistake about AI. They already have some talent in-house, and are quickly recruiting more. But the vast majority of newsrooms need to get smarter, better informed, and fast. Perhaps the best practical way to do this is to launch small experiments, using the new tools in such areas as graphics and illustrations, data analysis and versioning of already-produced stories, as well as on the business side. These sorts of steps may open minds to bigger possibilities.

a Sistine Chapel for our time?

Do we have the right mindset?

And opening minds is likely the highest priority. Trying to hold back the tide of change was a critical error last time around. Talk of “building moats,” simply accepting the limits of your own technological capabilities rather than making it a priority to expand them, or even avoiding steps that could free up resources because of staff resistance to cost-cutting, may be comforting in the short run, but these are all conservative steps in a moment when things are moving. This is no time to yield to an inclination to “stand athwart history, yelling Stop.”

One of the most important people in nurturing my own love of journalism was my high school print shop teacher, who served as adviser to the newspaper we set in type, printed and published. He came to teaching fairly late in his career, after his job as a newspaper linotypist was eliminated. In retrospect, it’s clear that those linotype jobs were eliminated not too quickly, but too slowly. The greatest days of newspaper publishing, both creatively and financially, came after they were all gone.

In sum, while this is likely a moment of anxiety for many in journalism who fear that technology is moving beyond them, it is in fact a time to build confidence and exercise agency. That will require education, experimentation and ultimately innovation.

How fast is fundamental change coming?

Having said that a revolution is soon upon us, that doesn’t mean that everything is changing everywhere all at once. One of the things I most noticed in looking back at the Nineties was that the technologists, and especially those with the arrogance to label themselves “futurists,” were often directionally right about change, but frequently far off in predicting its pace. Indeed, as I wrote,

it was often so long before they were right that they proved poor guides to people in business who needed to make decisions today… [T]o know where your industry is going is never enough in business. You also need to know when it is going to get there.

This dynamic may also apply in our own moment. In the last revolution, Xerox PARC, the early Apple (up through Newton), Knight Ridder’s Viewtron and Boulder Design Lab, Prodigy and AltaVista—brilliant innovators all--  proved to be ahead of their time. That may yet repeat itself with AI. The imperative, again, is therefore to become familiar with new tools, to bolster capabilities and to imagine a different future, rather than to bet the farm on someone’s particular speculative prediction.

What is the law—and what should it be?

As with the consumer internet in the Nineties, generative AI will be importantly shaped by the answers to a number of key legal questions, and by possible legislation. Here also, the major publishers are moving more quickly than they did last time around.

I will never forget a meeting of media lawyers I attended in the early years of this century. It featured endless complaints about online search, and whether it violated the legal rights of publishers, monetizing access to their content without offering them any compensation. By then, however, Google (the essentially monopoly provider) offered an opt-out (as it still does), which probably resolved the legal issue, but which none of the publishers believed they could afford to exercise.

Today, what remains of those same publishers are trying to negotiate with AI providers to license the right to mine their content. There are some early signs that a few of the larger publishers may accept deals in which they get paid, while most others are stiffed. Another case of the rich getting richer or, if you prefer, elites further entrenching themselves.

Perhaps even of greater moment is an emerging debate about the copyrightability of AI-generated content. To the extent that courts, or perhaps Congress, rule against copyrightability, the attractions of AI for publishers to create new material might be significantly reduced, as competitors could simply republish it. For those with teams of lawyers, including traditional trade associations and new ones like Rebuild Local News, this is likely an important fight just ahead.

It has been only nine months since the launch of ChatGPT and only seven since Microsoft’s investment in OpenAI rang the opening bell on the current speculative AI craze. Much has been said and written; lots of people have learned a good deal. It seems already clear that we are in the early days of profound change. I find that exciting, and hope you do too. This is a time for action, yes, but also, and even more importantly, for clear and careful thinking. I deeply appreciate that we get to do that together.