object(WP_Post)#7727 (24) { ["ID"]=> int(2172) ["post_author"]=> string(2) "36" ["post_date"]=> string(19) "2019-06-25 01:36:25" ["post_date_gmt"]=> string(19) "2019-06-25 01:36:25" ["post_content"]=> string(6573) "

The immediate challenge is not killer robots, its job replacement. If individuals are automated out of jobs, the future for society is bleak.

Computers can already take orders, fold clothes and even drive cars, but where to from here?

The robots are coming. Although often spoken of in future tense, the truth is machine learning is well and truly here. Without realising, consumers interact with ‘smart’ technology at almost every touch point; from robotic vacuums to facial recognition technology, artificial intelligence (AI) is helping to complete tasks faster, cheaper and – sometimes - more effectively than ever before.

In an economy that’s driven by speed and efficiency, it should come as no surprise that a computer’s ability to communicate at a trillion bits per second is favoured above the human capability of about 10 bits.

McKinsey recently reported that 40 per cent of work tasks can be automated using existing technology, prompting everyone from factory workers to lawyers and accountants to consider the threat of being replaced by robots as not just inevitable, but imminent.

For technologists, we are witnessing first-hand how this emerging field is transforming the companies we work for.

In my work at Isentia, we use machine learning to process seven million news items each day. Not long ago this was a task relegated performed solely by humans with the mind-numbing task of flipping through newspapers in search of stories that might relate to a client.

We have a duty to empower those around us to learn everything they can about what their job may evolve into in order to become the very best man-machine partner possible.

Today, machines trawl video, audio and digital content across over 5,500 new sites at a rate of 234 stories per second and present meaningful summaries to clients in real-time.

Whether a story breaks on Twitter and then spills across news platforms and onto television and radio, machine learning can track and analyse how a story evolves with 99 per cent accuracy.

While AI is revolutionising the way that we work, the impact is far greater for those in the tech industry. In our mission to develop software that can learn complex problems without needing to be taught how, the success of the AI industry ultimately comes down to technology professionals: our ability to automate, and the pace at which we expand the field of machine learning.

With an annual growth rate of 19.7 per cent percent (predicted to be worth $15.3 billion by 2019), it’s safe to say our foot is well and truly on the pedal. While this relies greatly on our technical capabilities, it is something that challenges many of us ethically: what set of values should AI be aligned with?

Two of the greatest technologists of our times, Elon Musk and Stephen Hawking, have spoken about both the potential benefit and the harm that an AI arms race could deliver. An eradication of disease is not unfathomable, but nor is a threat to humanity. They hold grave concerns as to whether or not robots can be controlled against misuse or malfunction.

While thought provoking, the immediate challenge is not killer robots, it’s job replacement. Employment may not seem like an ethical problem, but if individuals are automated out of jobs, the future for society is bleak. While the phrase ‘Thank God it’s Friday’ has forged its way into the 9-to-5 vernacular, for most people, jobs create a huge sense of personal and professional satisfaction… not to mention a means to pay bills.

An apocalypse might be somewhat melodramatic, however I do agree that it is important to consider just how closely we should merge biological and digital intelligence.

Computers can already take orders, fold clothes and even drive cars, but where to from here? It’s both exciting and terrifying. The last time we experienced a revolution like this was in the early 1900s when cars, telephones and the airplane all emerged at once.

Contrary to the hype, there lies an enormous opportunity for humans to work with artificial intelligence, not be replaced by it.

Make no mistake: at some level every job can be carried out by a robot. But there are certain jobs, particularly in technology, that require decision making, planning or coding software.

While computers do a brilliant job of executing well-defined activities - such as telling us the fastest route to get from home to work - it is safe to say that humans are an essential component of goal setting, interpreting results, humour, sarcasm and implementing common sense checks.

The most difficult jobs to automate are those that involve managing and developing people. While in this industry most of our jobs are safe (for now), we should heed the advice of Musk and Hawkings and protect those outside our field by proceeding with caution. How then to facilitate human and robots working together harmoniously without the workforce morphing into cyborgs? The secret is to not sail out farther we can row back.

As technologists, we also have a duty to empower those around us to learn everything they can about what their job may evolve into in order to become the very best man-machine partner possible. It's the best, and most ethical, way to prepare for the inevitable advent of AI.

First publish in CIO New Zealand

Andrea Walsh, CIO

" ["post_title"]=> string(39) "It's time to slow down the AI arms race" ["post_excerpt"]=> string(92) "Computers can already take orders, fold clothes and even drive cars, but where to from here?" ["post_status"]=> string(7) "publish" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(4) "open" ["post_password"]=> string(0) "" ["post_name"]=> string(38) "its-time-to-slow-down-the-ai-arms-race" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2019-06-26 01:01:55" ["post_modified_gmt"]=> string(19) "2019-06-26 01:01:55" ["post_content_filtered"]=> string(0) "" ["post_parent"]=> int(0) ["guid"]=> string(43) "https://isentiastaging.wpengine.com/?p=1829" ["menu_order"]=> int(0) ["post_type"]=> string(4) "post" ["post_mime_type"]=> string(0) "" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" }

object(WP_Post)#9089 (24) { ["ID"]=> int(49871) ["post_author"]=> string(2) "75" ["post_date"]=> string(19) "2026-09-09 02:19:40" ["post_date_gmt"]=> string(19) "2026-09-09 02:19:40" ["post_content"]=> string(17112) "

Audiences are no longer finding information through traditional search engines that favour established news outlets. AI models now highlight highly relevant and contextual information to audiences to often include niche and regional publications alongside major news media. This change challenges the old media hierarchy around tiered publications and pushes organisations to reconsider how and where they need to show up to stay visible in an AI-first world. 

Yes, organisations must focus on optimising their own content for LLMs, but will that always drastically increase the chances of AI models picking up your page? Probably not always. Smart strategy means targeting the specific publications your actual target audience reads — because those are the sources AI models retrieve when answering niche questions. 

It’s closer to digital PR than SEO

Generative Engine Optimization (GEO) is changing how brands approach online visibility. For years, traditional SEO meant focusing on your own site—optimising keywords, building backlinks, and improving on-page content. But AI models work differently. Instead of just using your website, these AI engines rely on trusted third-party sources to answer questions. This shift is taking place  gradually, of course. LLMs increasingly source from earned media (where it is accessible) and even offsite links from trusted sites. Owned media is still where the organisation has maximum control of how it’s own content travels, but a pivotal strategy shift is needed to match what AI models are picking up and citing.

To succeed with AI search, comms professionals need to think more like a digital PR strategist than a SEO expert. The best way to stand out is by earning mentions, quotes, and citations in the external publications your audience—and the AI systems they use—trust most. This does not make a distinction between Tier 1 or Tier 2 media. If AI models are crawling sites that mention an organisation, but the organisation does not acknowledge or even know those sites are being prioritised by LLMs, they risk falling behind in being the right kind of visible. 

To make this strategy work, looking beyond common metrics like traffic to the site or domain authority is not enough. Even a respected industry site will probably not influence AI answers as much if its content is behind a paywall or blocked from search engines. For AI visibility, accessibility to the site or page, structured data that can be crawled, and strong audience alignment are important. Since AI systems use both slow training cycles and fast real-time web searches (RAG), being featured on accessible, relevant niche sites helps an organisation show up accurately when models learn and when they search the web in real time. 

Why is Tier 2 media punching at Tier 1 weight?

According to Isentia's report How AI is destabilising trust and reputation amongst audiences, LLMs cite industry and trade publications about twice as often as traditional news sources. Company content and industry press make up over 60% of the share of voice LLMs use, while traditional news is twice as likely to generate negative sentiment. Thus, tier 1 outlets no longer automatically dominate AI-generated responses and may sometimes have the opposite effect.

Two main factors are driving this shift in which media is picked up by LLMs:

• The paywalled problem was further expanded on by Dr Momoko Fujita during the Digital News Report: Australia webinar that news organisations must figure out how to make paywalled content easily readable by LLMs. By bridging this gap, these organisations can ensure that AI tools deliver accurate, high-quality reporting rather than missing out on premium content. If not, high-quality coverage may never reach the model. Isentia’s Prashant Saxena, VP of Revenue and Insights, SEA, during a recent partner event with IABC APAC on Why AI Visibility is the next reputation frontier illustrated a paywalled Bloomberg story, for example, that was accurately summarised details it could read at the top level, but fabricated details about raised guidance, even though guidance had been cut. This is because it could not read the rest of the article and tried its best to assume what it can with the information that’s accessible.
  
• Specificity outweighs prestige. Tier 2, trade, and specialist publications are often more accessible, focused, and likely to provide the concrete, citable facts models need. Amy Chappell, Vuelio's Head of Insights Strategy, found a similar trend across sectors in her report on the visibility of supermarkets in the UK “ The role of AI, LLMs, and earned media in shaping reputation” and noted that supermarkets were most often cited by trade publications like The Grocer and Grocery Gazette, not national newspapers. Trade press stories, being more focused and well-sourced, provide models with clearer, more citable facts than broader national articles. This doesn’t mean that Tier 1 coverage does not matter — CEOs value front-page exposure because it remains highly influential. However, relying only on tier 1 hits now means missing significant AI visibility opportunities.

Cited vs consulted: LLMs read a hundred sources, but cite only a few

Which type of media gets cited relies upon how AI models scan different pages. If these models are citing much more niche media outlets, we can assume that a lot of these pages that are consulted could be a part of very relevant Tier 2 media that ends up actually getting cited, and that we’re seeing more and more examples of in AI answers.  At the IABC APAC and Isentia webinar on measuring brand visibility in AI answers, Prashant Saxena, Isentia's VP of Revenue and Insights for SEA, stated that in the search era "we would get sources on our page one, page two, mostly page one", and people would click through to form their own opinions. The combined click-through rate in that era was 35 to 40 per cent. Nowadays, he says, "it's just four to five per cent" — since LLMs provide a smooth, ready-made answer and "most of us aren't really checking the citations".

Communications teams now face a new consideration: the distinction between sources that are consulted and those that are cited. At the IABC APAC and Isentia webinar, Takeo Apitzsch, Hoffman Agency’s Chief Digital and AI Officer, explained that AI models scan hundreds of pages to generate an answer but cite only a select few to users. This means that the audience sees only a small, curated portion of the sources that actually influenced the AI's response and a lot of what actually shapes the AI answer doesn’t get visible credit. Therefore, organisations need to make sure they reach out to those publications that AI models can actually crawl and audiences trust the most. 

What does this mean for communications professionals?

We are seeing four practical shifts:

• Rebuild your tier list based on what LLMs actually cite, not on internal assumptions. A so-called “low-priority” trade publication or niche forum may contribute more to your AI visibility than a national outlet you have long targeted.
• Keep your reshuffled tier list fresh, not just correctly ranked. In Why is content freshness the new currency for AI visibility? we discuss that a page that hasn't been updated in eighteen months is far more likely to drop out of AI answers altogether, no matter how well it once performed. Getting the right tier 2 outlets on side is only half the job done. Feeding them (and your own owned channels) on an ongoing basis is the other half.
• Treat consistency as an essential. The largest gap between an organisation’s claims and what an LLM will confidently state is often due to inconsistencies between owned content and third-party coverage. When this occurs, the model may stop providing factual answers altogether.
• Shift your focus from share of voice to share of mind. It is now less about how much you are discussed and more about whether the systems mediating the most have got the correct information about your organisation.If the system holds the wrong version, your audience may never access the right one.

Structurally, as Ashley Knapp, Head of Brand and Corporate Affairs, East Asia at Schneider Electric noted during the webinar, these efforts can no longer remain siloed. Owned, earned, shared, and paid media have traditionally been managed by separate teams. Now, because of AI visibility, this required a unified approach, as models do not distinguish between departments but are first to detect inconsistencies. 

This also means reconsidering the PESO (paid, earned, shared and owned) strategy deployed by organisations since the way that LLMs access and prioritise them has changed. They prioritise brevity in content due to the high costs of GPUs and data centres. As a result, the shortest, clearest, and most trusted answers are favoured which benefits brands with strong reputations. Earned media remains important, but its influence now depends more on the credibility of the analyst than the platform. Shared content amplifies messages more than ever but is also where misinformation spreads fastest. Paid media is becoming more prominent in some models, though brands are still learning how this impacts visibility.

Media monitoring companies are becoming strategic AI visibility consultants

This shift requires media monitoring companies to evolve. Tracking mentions and sentiment across media channels has been central to media intelligence, but AI visibility has added a new dimension to this.  This means monitoring not only what is said about an organisation, but also which sources AI models use when answering questions about that organisation, and assessing how current, authoritative, and consistent those sources are. This gives media monitoring organisations an opportunity to own what they’ve developed and also be thought leaders in this space. Stakeholders value the “so what” advice much more than just knowing “this is what is being said about you in the media”.

Lumina AI View addresses this by tracking which sources ChatGPT, Gemini, Claude, and other models cite when representing an organisation, benchmarks citations against competitors, identifies narrative shifts before they reach stakeholders, and regularly scores AI visibility against four reputation pillars: Direction, Performance, Integrity, and Innovation, These pillars have always supported reputation management, now applied to a largely unseen audience.

If you're weighing up where a tool like this sits alongside the rest of your stack, our own comparison, Best AI Tools for PR & Comms Teams (2026), breaks down how AI-assisted coverage, measurement, crisis response and reporting tools stack up, Lumina included.

Because that’s really the mindset shift comms teams, and the firms advising them both need to make. As Takeo put it on the IABC APAC and Isentia webinar: “I fear that this is the mindset shift communications teams and their advisors must adopt. I fear that AIs will be your secondary, and if not, at least equal… audience in the future.” Beyond human visibility, reputation is about being accurately represented by the systems that mediate access to your audience, which is an additional layer that cannot be trivialised anymore. 

---

Want to see which sources are shaping how AI describes your organisation? Get in touch about Lumina AI View.

" ["post_title"]=> string(97) "How relevant is Tier 1 and Tier 2 media hierarchy in impacting how organisations show up in LLMs?" ["post_excerpt"]=> string(243) "The hierarchy that exists between Tier 1 & 2 publications today is being challenged. AI models are the new way audiences discover information requiring organisations to rethink how they show up to remain visible in an AI-mediated environment. " ["post_status"]=> string(7) "publish" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(4) "open" ["post_password"]=> string(0) "" ["post_name"]=> string(96) "how-relevant-is-tier-1-and-tier-2-media-hierarchy-in-impacting-how-organisations-show-up-in-llms" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2026-09-09 02:24:17" ["post_modified_gmt"]=> string(19) "2026-09-09 02:24:17" ["post_content_filtered"]=> string(0) "" ["post_parent"]=> int(0) ["guid"]=> string(32) "https://www.isentia.com/?p=49871" ["menu_order"]=> int(0) ["post_type"]=> string(4) "post" ["post_mime_type"]=> string(0) "" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" }

object(WP_Post)#12061 (24) { ["ID"]=> int(49595) ["post_author"]=> string(2) "75" ["post_date"]=> string(19) "2026-08-26 03:39:30" ["post_date_gmt"]=> string(19) "2026-08-26 03:39:30" ["post_content"]=> string(3210) "

---

" ["post_title"]=> string(63) "How AI is destabilising trust and reputation amongst audiences?" ["post_excerpt"]=> string(144) "Learn how LLMs reshape brand perception and actionable steps organisations can take to maintain trust and reputation in the new information era." ["post_status"]=> string(7) "publish" ["comment_status"]=> string(4) "open" ["ping_status"]=> string(4) "open" ["post_password"]=> string(0) "" ["post_name"]=> string(62) "how-ai-is-destabilising-trust-and-reputation-amongst-audiences" ["to_ping"]=> string(0) "" ["pinged"]=> string(0) "" ["post_modified"]=> string(19) "2026-08-26 03:46:01" ["post_modified_gmt"]=> string(19) "2026-08-26 03:46:01" ["post_content_filtered"]=> string(0) "" ["post_parent"]=> int(0) ["guid"]=> string(32) "https://www.isentia.com/?p=49595" ["menu_order"]=> int(0) ["post_type"]=> string(4) "post" ["post_mime_type"]=> string(0) "" ["comment_count"]=> string(1) "0" ["filter"]=> string(3) "raw" }