Gradually, Then Suddenly

We are in the second year of the current Generative AI cycle, and it's time for a reality check. Is this just another overhyped technology, or will it be transformational like the internet revolution a few decades ago? And who is really using it? What benefits are they seeing? AI really took off in the public consciousness with OpenAI's launch of ChatGPT at the end of 2022. Almost instantly, it set off a familiar technology cycle of substantial investment, considerable hype, and, inevitably, some level of disillusionment. So with regard to the professional publishers and communicators that Symbolic is focused on, where do we stand now? Generative AI, in its current form, is a classic technology platform. It is not naturally a button, a feature, or an add-on, but a foundational technology that enables an new ecosystem to emerge. Lots of new and well-established technology companies are now building applications on top of the Large Language Models with specific use-cases in mind. Symbolic.ai is focused specifically on the application of AI to the news, corporate communication and PR sectors globally. At the dawn of the internet (another platform), there was skepticism and slow corporate adoption. Initially, only the infrastructure providers were making money (Cisco was the world's most valuable company in 2000). Many enterprises created committees, studied the technology, and did small scale experimentation. But they were more cautious than opportunistic, and they did not play a significant role in shaping how the technology and its associated business models would develop. The internet turned out to be transformational, and many executives that I have spoken to privately regret that their businesses were passengers rather than drivers. For some industries it did not matter much, but other industries like news saw their businesses disrupted and their profit pools slashed.?

This Time, It's Different

Throughout the past year, Symbolic has engaged extensively with a wide variety of publishers to understand their perspectives on AI implementation. What we have seen and learned suggests that the pace of acceptance and adoption of AI is noticeably different from what was seen several decades ago as the internet emerged on corporate agendas.

A year ago, most companies we engaged with recognized AI as relevant but saw it as peripheral to core operations, using it if at all for tasks like post-production packaging. There was interest, but AI was rarely a top issue.

Six months ago, publishers began to shift their perspective, recognizing AI as a significant strategic issue. Our discussions frequently touched on threats, such as safeguarding and monetizing content in an AI world, as well as opportunities, including the internal automation of production tasks. They progressively started incorporating AI into essential processes like research.

Now, AI is a top issue for nearly all professional publishers. A flurry of licensing deals, lawsuits and business model changes is addressing the perceived threats to intellectual property, while the usage of AI internally is growing and increasingly impacting core production activities. The industry started slowly, but is now moving rapidly, asserting more control over AI's trajectory.

The publishers we talk to are now leveraging AI for research, document organization, data analysis, headline generation, social media content, and lightweight distribution. Though adoption is slower in core writing functions, this is shifting. AI increasingly supports the creation of newsletters, summaries, and derivative articles. And the cases of AI supporting original writing are increasing. This fast change from cautious observation to active deployment suggests that within the next year, publishers will need to have AI integrated across their operations to be competitive.

What Happens Next?

It's risky to predict how any new and important technology will evolve and grow, but there are a few principals that are worth betting on:

1) AI will not on its own be able to replace human writing other than for high volume low stakes publishing. AI will improve, but for a considerable time it will be a writing tool rather than a writer. Professional writers will remain in control, but they will need to understand how to use AI as part of their process.

2) It will become impossible for publishers to compete without using AI extensively across their technology stack. Just as no publisher would consider using a typewriter for its core functions, we are not far away from standard internet search, CMSs, and word processors being equally obsolete.

3) AI applications will become increasingly important as they verticalize and focus on specific company and industry use cases. Addressing narrow but important use cases and providing high touch service will continue to be critical in an AI world, even as the foundational models become more and more capable.

4) Publishers will need to be bold about their business models in the face of AI. AI will change subscriptions, advertising, and the click-based internet economy fundamentally. Playing defense is not a viable option in the face of this scale of change. Monetizing content as AI proliferates is an entirely new and different strategic proposition. This is an important topic that will be the subject of a future note.

— Devin Wenig

Symbolic Co-Founder and CEO