At the end of last year, the New York Times launched a high-profile lawsuit against OpenAI and Microsoft alleging copyright infringement and asserting that millions of pieces of content were used to train ChatGPT with permission. The law suit states:
While defendants engaged in widescale copying from many sources, they gave Times content particular emphasis when building their LLMs — revealing a preference that recognizes the value of those works. Through Microsoft’s Bing Chat (recently rebranded as “Copilot”) and OpenAI’s ChatGPT, defendants seek to free-ride on The Times’ massive investment in its journalism by using it to build substitutive products without permission or payment.
The case has yet to be resolved, but whatever the outcome, it’s likely to set an important precedent. Unease among content creators about generative AI ‘scraping’ their intellectual property (IP) in order to build Large Language Models (LLMs) for tech vendors was on the rise throughout 2023 as the gen AI hype built to near hysterical levels.
Theft of IP has been a common problem for publishers since content went online in a mainstream fashion. Here at diginomica, we were astounded a few years back to find a third party SAP consultancy business nakedly lifting our articles and publishing them to its customers/prospects. (When we took action, the CEO called us ‘assholes’!)
The perception that if it’s on the internet, it’s fair game for lifting and sharing hasn’t gone away, but the rise of gen AI has upped the stakes significantly. As The Times puts it in its suit:
If The Times and other news organizations cannot produce and protect their independent journalism, there will be a vacuum that no computer or Artificial Intelligence can fill.
In other words, this matters and a line in the sand needs to be drawn.
Flash forward to this week and CEO Meredith Kopit Levien is limited in terms of what she can say in direct relation to the suit, but is quick to point out that The Times is not closed to doing partnership deals with AI tech providers:
You can imagine we are talking to potential partners all the time. And you've seen directly in the complaint that we're talking of potential generative AI partners. I'd say we are being really selective and thoughtful about what partnership makes sense for us in the context of our strategy and our rights being respected to our IP and fair value exchange for that IP.
A useful exemplar also came in December when The New York Times Company signed a multi-year content licensing deal was signed with Apple News Plus for particular assets. Kopit Levien explains that there are a number of elements that the publishing giant looks for in such agreements:
Does this make sense in the context of our essential subscription strategy, by which I mean, is it helping us in some way to build direct relationships? That can be through very direct ways or through helping us grow audience and awareness for our brand. In the case of this partnership we've just built with Apple News Plus…we certainly see it as doing that.
And then we look at the dimension of our IP rights being appropriately respected and used in a responsible way and is there fair value exchange? I would say any time we look to engage with a big tech partner - or any partner for that matter - those are the considerations we're making. And I think we've got a good track record of doing deals that live up to those standards
The most important thing we can do as a business…is to have products that are so valuable at scale and widely understood for their brand marks and the credibility and the trustworthiness they provide, that people are inclined to build a direct relationship with us, and will come to us on a very regular basis. So that's the main game we are playing, to the extent that there are other ways to tap into the vast arsenal of IP that we create every day that we've created for close to two centuries. We are open to that as long as there is fair value exchange and it's supportive of the broader business model.
Wooing, not suing
Over at News Corp, CEO Robert Thompson wants to take a partnership approach with gen AI firms:
While certain other media companies prefer litigation, we prefer consultation, as the former is merely creating a gold rush for lawyers. Courtship is preferable to courtrooms. We are wooing, not suing.
Artificial Intelligence, with all its permutations and perturbations, will play an increasingly important role at most businesses. We expect to be a core content provider for generative AI companies who need the highest quality, timely content to ensure the relevance of their products. The corny cliché is that AI companies are selling the picks and shovels during this seeming gold rush. Well, we are selectively reselling gold nuggets and those crucial negotiations are at an advanced stage.
News Corp has led the debate around AI and IP, he boasts, and is in no doubt about the possible downsides:
I mean, 17 years ago, when prestige-craving media executives were sashaying with Silicon Valley, we were raising doubts, doubts about provenance, but also about the baleful impact on vulnerable young people, the smartest engineers on the planet, creating compulsive, addictive experiences.
We’re certainly not naive as well about the potentially positive and negative impacts of AI on our journalism, our creativity, our content. We’ve had almost two decades of distribution dominating creation…but gen AI is a hyper-effective form of derivative distribution. It’s retrospective, not prospective, and the thoughtful AI companies understand that fact…Thoughtful people do understand that counterfeiting is not creating, and crucially, in this exceedingly erratic era, we have deep facts, not deep fakes.
Among such thoughtful firms, Thompson ranks OpenAI and its CEO Sam Altman, who has “shown a clear understanding of the social importance of journalism”. Others are less aware:
In my view, those who are repurposing our content without approval are stealing. They are undermining creativity. Counterfeiting is not creating and the AI world is replete with content counterfeiters.
In terms of the wider debate, Thompson has a clear set of goals in mind:
We are hopeful that, again, News Corp will be able to set meaningful global precedents with digital companies that will assist journalists and journalism, and ensure that gen AI is not fueled by digital dross. We speak of the AI hallucinating, yet we as a society are hallucinating if we don’t focus firmly on provenance at a time when even the very words ‘misinformation’ and ‘disinformation’ have themselves become sources of misinformation and disinformation.
Thompson has a warning shot for other media companies about their role in shaping the future:
Too many media companies are scanning the landscape and presuming that they have a glimpse of the future, and yet they cannot distinguish between trendiness and actual trends. Too many media companies, for too long, have been guilty of the Abilene Paradox.
The Abilene Paradox is a management theory first outlined by Jerry B. Harvey in 1974. It is a collective fallacy whereby a group of people decide on a course of action that is actually counter to the good of all the individuals in the group. But ‘groupthink’ prevails and results in potential outcomes that no-one really wants.
Frankly, a ruling in the New York Times Company’s suit can’t come soon enough. In the meantime, watching which tech firms are willing to come to mutually-beneficial deals with content creators is going to be revelatory in its own right.