Dominic Young, Chief Executive Officer/Founder, Axatewritten evidence (FON0021)

 

House of Lords Communications and Digital Select Committee inquiry: The future of news: impartiality, trust, and technology

 

 

1                        Summary and proposed recommendations

 

News media decline is accelerating.

 

1.1.1         The decades-long decline in the profitability of news media has accelerated rapidly in the last few months. Large scale job losses are being driven by the reality that few, if any, news media businesses are achieving sustainable revenue growth.[1] They face a future of continued disruption and decline and are configuring their resources accordingly. Some commentators suggest this decline is now terminal; others assert that the only future for journalism will be as a non-profit activity, requiring wealthy patrons or state funding to continue.

 

This decline can be arrested and reversed.

 

1.1.2               My analysis of the reasons for this decline, and the impact of big tech platforms and AI, are below. Overall, though, my view is fundamentally more ambitious and optimistic than the gloomy commentary mentioned above. This is for two reasons.

 

              -               There is still a considerable un-met consumer need for professionally produced, journalistically driven, news and media. The failure of news media to thrive is a consequence of the digital environment rather than any inbuilt obsolescence.

 

              -               Additionally, as this Committee’s excellent report[2] on Large Language Models puts it, “Large language models (LLMs) will introduce epoch-defining changes comparable to the invention of the internet.” As organisations and legislators work together to shape this new epoch, we have the opportunity to reconsider and correct, with the benefit of hindsight, some of the missteps of the recent past. These led to the toxic online environment in which news media has struggled to survive.

 

The Committee’s recommendations can help define a positive future for news media.

1.1.3         This Committee’s strong support for copyright should be repeated in this inquiry. Copyright is a bedrock issue for the news media. Enforcing it, without special exceptions or evasion by any organisation, including the platforms, will restore and re-balance the necessary incentives for journalism.

 

1.1.4         I would like the Committee to support independent news media by making recommendations on the following lines:

 

1.1.5         A firm endorsement of copyright and its fundamental principles. Copying and exploitation of copyright works requires consent. New exceptions, especially for AI, are neither required for a healthy AI sector, nor desirable in principle. It would be very helpful for the Committee to say so.

 

1.1.6         Broad-based liability exceptions such as those in the European e-commerce directive,[3] and similar provisions in legislation such as the Digital Millennium Copyright Act[4] and Section 230 of the Communications Decency Act in the USA[5] for tech platforms have led to unintended consequences including their untrammelled growth and market dominance. Such exceptions, especially for text and data mining, should not apply to AI platforms and must not be extended to them. Additionally, the existing exceptions in UK law should be narrowed, making platforms more fully accountable for the content they publish. This will help restore a healthier online environment, and not just for news. It is also relevant to debates around other harms created by online platforms such as the very serious issues of child safety, online bullying, deepfakes and more. The Committee should back this reform.

 

1.1.7         This will also help address the issues of trust online, because platforms and/or their users being accountable for the content they publish, with full liability under the law, will remove their existing incentives to ignore the legality or accuracy of content in pursuit of readership or income. While nothing will stop completely the flood of disinformation and untruth on to the internet, increased accountability will at least mark out trusted publications and platforms in the eyes of consumers. Good money will drive out bad.

 

1.1.8         After decades of pressure and decline, the media will need help in rapidly adapting to the demands of the next epoch. The Committee should urge Government to consider providing assistance not in the form of subsidies or direct support, but creating standards and infrastructure to support the rapid emergence of a new norm online whereby the seeking and granting of permission is the default. Contributing proper funding for projects such as The Copyright Hub will help the establishment of a new marketplace for rights and permissions which will, in turn, be a significant enabler of new opportunities for news media.

 

1.1.9         I would be happy to give evidence in person to the Committee to explore some of the themes and ideas outlined in this written response.

 

2                        About the author

 

My expertise is in the field of copyright and IP, licensing, digital rights issues and digital media business models, particularly as these apply in news media. In this response I am speaking for myself, not any particular organisation or interest group.

 

I previously held a range of senior operational and strategic roles at News UK, including responsibility for its licensing business and intellectual property. I also undertook a number of roles representing the industry in rights negotiations and industry-wide initiatives such as the Automated Content Access Protocol.

 

My other roles have included leadership of The Copyright Hub between 2013-2016, a government-initiated project that aimed to simplify the process of content licensing. I was a director and chairman of NLA Media Access (Newspaper Licencing Agency) whose initiatives have generated around £750m for the UK news industry sector to date. I steered the development and launch of its eClips database, which created a centralised database of newspaper content which has been driven the NLA’s core digital cuttings service and used in numerous ways by and on behalf of the participating publishers. This project offers a number of positive lessons for anyone considering models for industry collaboration in the coming era of a rules-based internet.

 

I write[6] and speak extensively about the challenges facing the media and ways of addressing them. Several years ago, I founded a company called Axate[7] to address one of the main challenges facing the news industry today: a reader-friendly, publisher-centric solution to generate sustainable, reader-driven revenue and independence from the big platforms. It enables any user to casually pay for access to any participating digital publication on a per-article, per-day/week/month basis without having to make any commitment and multiple sign-ups. Our goal is to help bring regular, frequent and casual payment for media to the mass market. This directly addresses the biggest existential threat to news media, which is the absence of reliable, scalable revenue models.

 

My answers focus on digital news publications specifically, rather than on print newspapers or broadcasting.

 

 

 

3                        What impacts (positive and negative) do large technology platforms and online news aggregators have on the UK’s news environment, including media plurality? And how might this change?

 

Tech platforms dominate the internet and create perverse incentives which damage news media.

 

3.1.1         Online platforms largely dictate the commercial and editorial incentives in the news environment, especially for publishers that are dependent on advertising as their primary source of income, rather than reader revenue. The net effect of this has been overwhelmingly negative. The incentives they create have been at the heart of the ongoing decline of the news industry. Here are the reasons why:

 

The volume Incentive.

 

3.1.2         Large platforms - initially search engines and then social platforms - have incentivised news media to optimise their product offerings to position the large platforms as the primary stakeholder. SEO (Search Engine Optimisation) has become a core function in news organisations. News organisations for which advertising is the principal revenue source have to make achieving the highest possible volume of traffic a key target, which means they have to focus their attention on how to manipulate platform algorithms to maximise it.

 

3.1.3         This often produces, as you might have noticed from your experience reading advertising-dependent news sources online, consumer experiences for humans which have degraded over time, for example by slowing page loading times, adopting intrusive advertising formats and allowing advertising to take precedence over content. Algorithms are the key stakeholders, so content is optimised to generate the greatest algorithmic payback. Changes to these algorithms, or a failure to optimise for them can have a dramatic impact on news organisations.

 

The content/product incentive.

 

3.1.4         Big platforms don’t deal in news products; they deal in individual articles (or, more precisely, the data they can gather about users when they seek and access those articles). While it might be relatively easy to attract traffic to an article, it’s very hard for publishers to develop that into a longer-term relationship with a user, even more so if the user never clicks through to the publisher’s website or app. Few users choose to return to a news article they have enjoyed, to experience that enjoyment again and again. They do, however, return to products and brands which they enjoy and trust. But when publishers are reliant on platforms for much of their traffic, their incentives move away from trying to create a product which is satisfying and habit-forming as a whole, and towards creating large volumes of content cheaply, which will perform well in the algorithmic world of big platforms. This feeds the platforms’ need for more data, while driving some traffic back to the publisher.

 

The terms-of-trade incentive.

 

3.1.5         Ad-dependent websites have little control over the price and terms under which their advertising is sold. The need to maximise advertising revenue without strong enough influence over prices explains consumer experiences which can be less enjoyable or borderline hostile: pages covered in ads to the point of obstructing the user experience. The huge control which the big platforms have over the advertising market exacerbates the challenges of competing for ad revenue with a far greater competitive set. Put simply, it’s hard to have a good business without much control over your terms of trade.

 

These incentives undermine the needs of publishers and their audiences.

 

3.1.6         These incentives conspire to mean that consumer interests and needs are forced to take second place to commercial reality for many news publishers. As revenue declines over time,  commercial viability and ability to invest decline with it.

 

But if we change the incentives, we can change the outcome.

 

3.1.7         However, the reason I frame this in terms of incentives is because I believe that if we can change the incentives in the current market, we can also change the outcomes. Extinction is not inevitable.

 

3.1.8         It is not the free market that is driving news organisations to extinction. If it were, perhaps we could say that what’s happening now is just market forces at work. We could stand back and let the market determine the outcome.

 

3.1.9         But this is not a free market in the traditional sense. It is a competition between a large number of reputable publishers trying to provide a diversity of news, and a handful of unimaginably colossal global platforms which have manipulated the market to achieve total dominance of revenue.

 

3.1.10     Diversity and plurality of news is essential for a well-functioning society and community. Even publishers who strive for impartiality end up with strikingly different output. The history of the British newspaper industry, with each title’s attitude and partiality openly and proudly expressed in a highly competitive and plural market, underscores the point. A single voice like the BBC, no matter how respected, cannot supply that wonderful ecosystem of choice. Choice empowers citizens and fuels debate, and it needs to be protected.

 

4                        How is generative AI affecting news media business models and how might this evolve?

 

It’s too soon to say.

 

4.1.1         The short answer is: we don’t know yet. The whole subject of AI is characterised by feverish speculation about what will happen in the future. It’s a form of crystal ball gazing which we also saw around the start of the mass-market internet. Some stare into their crystal ball and see a utopian future, others the opposite. Everyone is scrambling to do something, without knowing exactly what, or even what outcome they might be hoping for.

 

Current GAI products offer no benefits for news media.

 

4.1.2         Focusing on general GAI products as they are now and putting aside for the moment the possible applications of AI technology within news organisations, we can see fairly clearly that there has so far been no benefit for the news industry at all.

 

The use of publishers’ content without permission is a massive copyright infringement.

 

4.1.3         News publishers’ content has been used on an industrial scale to train AI models which are designed to substitute for the sources of their training data.

 

4.1.4         Publishers have not been asked for permission: attempts to licence some content have begun belatedly and with only a select few publishers. There has been little transparency about what content has been used how, however legal actions such as that brought by the New York Times against OpenAI[8] include allegations, supported by examples, that copyright has been infringed both when training AI systems and, on occasion, in their output.

 

AI output can be seriously risky for news publishers and consumers.

 

4.1.5         Worst of all, the AI output is unreliable and potentially highly risky.

 

4.1.6         This risk is especially high in relation to news, where the propensity of GAI systems like ChatGPT to “hallucinate” means entirely made-up information may plausibly be presented as facts.

 

4.1.7         Hundreds of AI-generated and seemingly bona fide news publications are already up and running.[9]

 

4.1.8         The obvious risk is exacerbated for bona fide news publishers by the possibility that false content might become incorrectly associated with their brand (by attributing a source, for example) which would create a serious reputational risk.

 

4.1.9         As OpenAI put it in Section 2.2 of GPT 4’s “system card”[10]

 

GPT4 has the tendency to “hallucinate”, i.e. “produce content that is nonsensical or untruthful in relation to certain sources. This tendency can be particularly harmful as models become increasingly convincing and believable, leading to overreliance on them by users.

 

The implications of GAI are profound and we should not be afraid to challenge the status quo.

 

4.1.10     The issues raised by the advent of GAI are profound, perhaps more so than the current debate suggests. AI companies demand the right to freely use any content they find online to train their models. They say that without it they will be unable to continue their current trajectory.[11] We should be wary of taking that threat too seriously or even taking it as a threat at all to anyone other than their investors, given the numerous problems with the current reckless way in which GAI is being pursued.

 

4.1.11     If we are going to give AI companies, and their platform owners and patrons, the unfettered right to use and exploit anyone’s work, it will cause huge economic, cultural and creative destruction which we can ill afford. We have recent experience to look back on and learn from.

 

We should learn the lessons of the recent past and not repeat historic mistakes.

 

4.1.12     In an earlier era of the internet, concerns were raised about the potential of copyright and other laws to inhibit the development of a new generation of businesses emerging as the internet began its explosive growth. Search engines in particular were built with, as with AI now, a complete disregard for copyright. Good arguments were made. How can someone who is merely operating a server even know, far less be legally accountable for, whatever their users get up to?

 

4.1.13     Copyright, the boring but incredibly successful legal principle that you can’t use someone’s work without asking, was largely swept away along with numerous other legal liabilities which the internet platforms were largely given permission to ignore. That legal impunity rapidly led to the development and explosive growth of platforms which are now so large, powerful and valuable that they distort every online economy and opportunity.

 

4.1.14     Copyright’s long track record of creating huge, diverse, open and competitive marketplaces is a precious and powerful economic and cultural engine. When it’s undermined, as it was for the benefit of the big platforms mentioned above, it destroys value and creates monopolies. We must not repeat that historic mistake when it comes to AI.

 

5                        What factors affect trust in news and how might this evolve?

 

Accountability, legality and integrity, rather than just impartiality, are key to trust and will become more so as the volume of AI-generated content continues to surge.

 

Accountability.

 

5.1.1         Who published this? What’s their name? How do I contact them? Will their reputation and commercial fortunes suffer if they have got things wrong or if they spread incorrect information? Are they members of any self-regulatory standards bodies who can help address and correct any mistakes? Having an actual person or company standing accountable behind news content is an important element of trust. The same cannot be said for AI-generated content.

 

Legality.

 

5.1.2         Is this publisher fully accountable under the law? Could they be sued or prosecuted if they have broken any laws? Infringed any copyrights? Published something defamatory? It ought to go without saying, but the default online is that big platforms, and others, are largely immune from legal consequences. They deal with complaints by exception, taking content down long after it is posted and has ceased generating revenue. There is no pre-publication scrutiny of content, no fact checking, no verification of sources or copyright, no screening for rogue AI-generated content.

 

5.1.3         Being fully subject to the law distinguishes trustworthy publishers from others exactly because they undertake these responsibilities and are accountable for any mistakes. This needs to be highlighted to users who are concerned about trust.

 

Integrity.

 

5.1.4         What is this publication promising me, and are they living up to it? That might sound vague but it’s important. Many publications are focused on, and dedicated to, finding, checking and publishing the truth, however uncomfortable. They make that promise, they hold themselves to high standards and consumers should do likewise. Others are all about parody and satire. They don’t promise absolute truth, necessarily, but you can still trust them to deliver their promise.

 

5.1.5         I offer these observations to try to frame the question of trust. It’s not a simple, easily measured, thing. Especially in the context of a thriving, plural and competitive news market, where consumer trust is a key factor of success, these factors of accountability, legality and integrity will be important in distinguishing trustworthy sources from others.

 

 

6                        How well are news organisations responding to factors affecting their business models, and are any changes needed?

 

News media remains in long-term and accelerating decline.

 

6.1.1         News organisations have self-evidently failed to halt or reverse their long-term decline and continue to operate in the shadow of the dominant platforms. This has, among other things, stifled innovation within the media sector and reduced its attraction to investors and entrepreneurs.

 

Tech platforms are largely responsible.

 

6.1.2         The online environment can take a large share of the blame for this, and the role of the big platforms mentioned above is critical because they have monopolised advertising revenue and access to users, created perverse incentives and exploited content without permission or payment.

 

Legislators and regulators also share the blame.

 

6.1.3         Legislators have to take their share of the blame. Creating legal immunity for some parts of the online ecosystem, such as those in the European e-commerce directive,[12] the Digital Millennium Copyright Act[13] and Section 230 of the Communications Decency Act in the USA,[14] was a historic mistake and has resulted in a toxic online environment. These immunities should have been reduced, narrowed and limited when their harms became obvious, but little has been done.

 

6.1.4         Regulators, for their part, have been slow to respond to the growing dominance and unfettered expansion of the big tech platforms and have allowed, through inaction, that dominance to expand and metastasize.

 

News organisations must accept responsibility as well.

 

6.1.5         However, news organisations can’t completely escape responsibility for their own parlous state. Their business model has always depended on two sources of revenue, broadly summarised as advertising and consumer. With the advertising market permanently disrupted by many new competitors (as well as dominated by the platforms), they have failed, so far, to adopt better consumer revenue models and to evolve their product offerings to meet the needs of more of their audience.

 

 

 

Awaiting rescue is not going to be enough.

 

6.1.6         Some news organisations are, essentially, awaiting rescue, hoping that the metaphorical tide will turn, that pay-outs from big platforms will be forced by governments or lawsuits, or hoping that AI companies will pump billions of dollars into compensating them. Even if all this happens, and the appropriation of value from the news sector is reversed, that change on its own will only be the start of what the industry needs to thrive in the long term.

 

News organisations need to take control of their destinies.

 

6.1.7         To create a diverse, competitive, sustainable, pluralistic news industry, publishers need to take control of their own destinies. By becoming more reliant on consumer revenue, and focussing their products on the readers (rather than the incentives that platforms drive), publishers will gain more economic independence and ultimately can thrive even without any other changes to the online environment. It is within publishers’ power to make the change needed for a better future for the industry.

 

6.1.8         The news industry needs consumer revenue models which can work for every publisher and, crucially, every consumer. Subscription models are successful for a very small proportion of the audiences of an even smaller group of news publishers. Each of them has built its own shop, which only sells its own product and will only sell it to people who are willing to continuously pay for it. Finding new subscribers, and keeping existing ones, is a challenge which eventually starts to be lost.

 

6.1.9         There is no such thing, without solutions like Axate, as a spontaneous, casual sale of the kind we are used to in newsagents in the real world. Until there is, consumer revenue will remain a niche model in the news sector, limited to affluent and highly committed readers.

 

6.1.10     However the need and demand for news remains near-universal. With advertising unable to support the sector, there’s a real opportunity to meet the needs of the mass market in a new way. The industry needs to act, not wait for others to do so for them.

 

7                        How adequately are UK news organisations providing impartial and trusted news? What actions are needed to address any shortcomings?

 

Impartiality requires plurality.

 

7.1.1         As noted above, it’s hard for any one news organisation to be genuinely impartial, and impartiality is not the role of news publishers outside of regulated public services. The act of editing, selecting what to cover and what not to, is not neutral. Even if you endeavour to focus reporting on the facts, they will always be the facts as seen by the reporter, selected by the editor. Others will see and report something slightly different. The whole point of commercial news media is to provide a diversity of opinion and perspective across a range of publications which are available to all who choose them.

 

7.1.2         The best answer to concerns about impartiality and trust is to ensure that the market supports a wide range of trusted publications, that it is open to new competitors, that the consumer is the key stakeholder in their business models and that everyone can access everything if they choose.

 

7.1.3         Right now, we are doing a terrible job on that front. Some of the most trusted content is restricted to subscribers only, who are a tiny percentage of the readership. We need to fix this, and we need better commercial incentives for mass-market high-quality journalism.

 

7.1.4         We need to fix the market for that situation to change.

 

 

8 February 2024

11


 


[1]              https://www.reuters.com/business/media-telecom/daily-mirror-parent-reach-plans-cut-450-full-time-jobs-2023-11-08/; https://www.politico.com/news/2024/01/23/los-angeles-times-cuts-newsroom-washington-00137301; https://www.nytimes.com/2024/01/19/business/media/sports-illustrated-mass-layoffs.html; https://www.nytimes.com/2024/01/17/business/media/music-magazine-pitchfork-merger-gq.html;

[2]              https://publications.parliament.uk/pa/ld5804/ldselect/ldcomm/54/5403.htm

[3]              https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=celex%3A32000L0031; https://assets.publishing.service.gov.uk/media/5a758afced915d506ee7fa20/livelink-livelink-livelink.exe-43464131-eCommerce-guidance.rtf-funcdoc.....pdf

[4]              https://www.govinfo.gov/content/pkg/PLAW-105publ304/pdf/PLAW-105publ304.pdf; https://www.copyright.gov/dmca/

[5]              https://www.govinfo.gov/app/details/USCODE-2021-title47/USCODE-2021-title47-chap5-subchapII-partI-sec230

[6]              https://medium.com/@dominic_young

[7]              https://www.axate.com

[8]              https://www.nytimes.com/2023/12/27/business/media/new-york-times-open-ai-microsoft-lawsuit.html?unlocked_article_code=1.U00.WCOG.Ak3z95Gf2BAr&smid=url-share

[9]              https://www.newsguardtech.com/special-reports/newsbots-ai-generated-news-websites-proliferating/

[10]              https://cdn.openai.com/papers/gpt-4-system-card.pdf

[11]              https://www.theguardian.com/technology/2024/jan/08/ai-tools-chatgpt-copyrighted-material-openai

[12]              https://eur-lex.europa.eu/legal-content/EN/ALL/?uri=celex%3A32000L0031; https://assets.publishing.service.gov.uk/media/5a758afced915d506ee7fa20/livelink-livelink-livelink.exe-43464131-eCommerce-guidance.rtf-funcdoc.....pdf

[13]              https://www.govinfo.gov/content/pkg/PLAW-105publ304/pdf/PLAW-105publ304.pdf; https://www.copyright.gov/dmca/

[14]              https://www.govinfo.gov/app/details/USCODE-2021-title47/USCODE-2021-title47-chap5-subchapII-partI-sec230