<p><em><strong>Digital Disconnect:</strong> </em>For over a decade, the technology industry has moved at breakneck speed under a familiar justification: innovation first, guardrails later. But something unusual is happening inside the artificial intelligence ecosystem. The warnings are no longer coming only from regulators or privacy activists. They are coming from the very researchers and co-founders who helped build these systems.</p>
<p>And that shift matters.</p>
<h3><span style=”color: #ba372a;”><strong>When Intimacy Meets Advertising</strong></span></h3>
<p>At the centre of the latest debate is OpenAI and its flagship product, ChatGPT.</p>
<p>Former OpenAI researcher Zoe Hitzig recently cautioned against integrating advertising into ChatGPT, arguing that the chatbot has accumulated something unprecedented: an archive of deeply personal human disclosures.</p>
<blockquote class=”twitter-tweet”>
<p dir=”ltr” lang=”en”>“OpenAI Is Making the Mistakes Facebook Made. I Quit.”<br /><br />“…Many people frame the problem of funding A.I. as choosing the lesser of two evils: restrict access to transformative technology to a select group of people wealthy enough to pay for it, or accept advertisements even if it… <a href=”https://t.co/h1BqS6linA”>pic.twitter.com/h1BqS6linA</a></p>
— kristen shaughnessy (@kshaughnessy2) <a href=”https://twitter.com/kshaughnessy2/status/2021650823671926918?ref_src=twsrc%5Etfw”>February 11, 2026</a></blockquote>
<p>
<script src=”https://platform.twitter.com/widgets.js” async=”” charset=”utf-8″></script>
</p>
<p>Unlike social media platforms, where users curate public personas, AI chats are often unfiltered. People discuss medical anxieties, financial distress, relationship breakdowns, religious doubts, and career dilemmas with a candour rarely seen elsewhere online. That intimacy, she argued, fundamentally alters the stakes of monetisation.</p>
<p>OpenAI has stated that it does not sell chat data to advertisers and that ads will not appear alongside sensitive health or political discussions. It maintains that conversations remain private. The concern, however, is not solely about current safeguards. It is about incentives.</p>
<p>Ask any digital advertising expert, and they’d agree that the sector historically rewards engagement. The longer users stay, the more opportunities exist to monetise attention. While OpenAI has said it does not design ChatGPT to maximise engagement, critics note that such commitments are voluntary, not legally binding.</p>
<p>The deeper question is whether a tool positioned as a digital assistant, tutor, and quasi-confidant can ethically coexist with targeted advertising, even if that targeting is based only on broad topical signals. When commercial pressure increases, structural priorities often shift incrementally rather than abruptly.</p>
<p>The risk is not an obvious banner ad next to a health query. The risk is subtle behavioural optimisation.</p>
<h3><span style=”color: #ba372a;”><strong>Internal Fractures At xAI, Anthropic</strong></span></h3>
<p>Parallel turbulence is visible at xAI, founded by Elon Musk. Two of its co-founders recently resigned, reducing the original leadership team significantly in less than three years. Reports have suggested tensions over performance demands and the pressure to close the competitive gap with rivals such as OpenAI and Anthropic.</p>
<blockquote class=”twitter-tweet”>
<p dir=”ltr” lang=”en”>I resigned from xAI today.<br /><br />This company – and the family we became – will stay with me forever. I will deeply miss the people, the warrooms, and all those battles we have fought together.<br /><br />It’s time for my next chapter. It is an era with full possibilities: a small team armed…</p>
— Yuhuai (Tony) Wu (@Yuhu_ai_) <a href=”https://twitter.com/Yuhu_ai_/status/2021113745024614671?ref_src=twsrc%5Etfw”>February 10, 2026</a></blockquote>
<p>
<script src=”https://platform.twitter.com/widgets.js” async=”” charset=”utf-8″></script>
</p>
<p>The company’s chatbot, Grok, has also faced controversy over ‘undressing’ people. Broader generative AI ecosystems have struggled with misuse cases, including non-consensual image manipulation, where AI systems are exploited to create fabricated explicit images of women.</p>
<p><strong>ALSO READ: <a title=”Elon Musk’s Grok AI Is Still Undressing Men” href=”https://news.abplive.com/technology/grok-ai-undressing-people-men-nude-controversy-explained-1825614#google_vignette” target=”_self”>Elon Musk’s Grok AI Is Still Undressing Men</a></strong></p>
<p>Even in Anthropic, the $380-billion powerhouse behind Claude, AI safety researcher Mrinank Sharma resigned with a stark warning to others: “The world is in peril.” </p>
<blockquote class=”twitter-tweet”>
<p dir=”ltr” lang=”en”>Today is my last day at Anthropic. I resigned.<br /><br />Here is the letter I shared with my colleagues, explaining my decision. <a href=”https://t.co/Qe4QyAFmxL”>pic.twitter.com/Qe4QyAFmxL</a></p>
— mrinank (@MrinankSharma) <a href=”https://twitter.com/MrinankSharma/status/2020881722003583421?ref_src=twsrc%5Etfw”>February 9, 2026</a></blockquote>
<p>
<script src=”https://platform.twitter.com/widgets.js” async=”” charset=”utf-8″></script>
</p>
<p>Take a minute to let that sink in. A PhD holder in Statistical Machine Learning from the University of Oxford, who was in charge of a team working on AI guardrails, decided to move away, pursue his love for poetry, and “become invisible.”</p>
<h3><span style=”color: #ba372a;”><strong>Release, Observe Misuse, Patch, Repeat</strong></span></h3>
<p>While companies typically frame such outputs as user misuse or guardrail failures, critics argue that these risks are intrinsic to models trained on vast, imperfect internet datasets. When realism is prioritised without equally robust safety architecture, harmful edge cases become predictable outcomes rather than anomalies.</p>
<p>The industry’s pattern has been reactive: <em>release, observe misuse, patch, repeat. </em>That approach may be insufficient as systems scale and become embedded in everyday life.</p>
<h3><span style=”color: #ba372a;”><strong>AGI Narrative & Scientific Scepticism</strong></span></h3>
<p>Beyond advertising and misuse controversies lies a larger intellectual dispute. Many AI executives frame their work as progress toward artificial general intelligence, or AGI, a system capable of matching or exceeding human cognitive ability across tasks.</p>
<p>Yet not all researchers agree that scaling large language models (LLMs) inevitably leads to general intelligence. Cognitive scientists have pointed out that modelling language patterns does not equate to replicating human reasoning, which involves perception, embodiment, emotion, and context beyond linguistic prediction.</p>
<p>This scientific caution clashes with corporate rhetoric. AGI, as a concept, justifies extraordinary investment, vast data centre expansion, and substantial environmental cost. It also fuels competitive urgency.</p>
<p>If the destination is transformative intelligence that solves humanity’s hardest problems, then aggressive development appears defensible.</p>
<p>But if the underlying architecture cannot realistically deliver that outcome, the moral calculus changes.</p>
<h3><span style=”color: #ba372a;”><strong>Incentives, Not Intentions</strong></span></h3>
<p>It is important to note that none of the departing researchers has accused their companies of immediate wrongdoing in the contexts described. The anxiety centres on the trajectory.</p>
<p>AI is still a Wild, Wild West area, primarily thanks to the lack of strict policies globally. Technology companies, hence, often begin with principled commitments. Over time, however, business models exert gravitational pull. Engagement metrics influence design. Revenue pressures reshape priorities. Boundaries that once seemed firm become flexible.</p>
<p>History provides precedent.</p>
<p>Social media platforms like Instagram or TikTok initially promised connection and openness. They later grappled with misinformation, addiction dynamics, and data exploitation. AI systems operate at an even deeper psychological layer. They are interactive, responsive, and increasingly integrated into personal decision-making.</p>
<p>If advertising, engagement optimisation, or competitive pressure become dominant drivers, the consequences may not resemble past platform harms. They may be more intimate and more opaque.</p>
<h3><span style=”color: #ba372a;”><strong>A Turning Point Or A Familiar Cycle?</strong></span></h3>
<p>The departures at OpenAI and xAI do not signal collapse. Both companies remain financially powerful and technologically influential. But dissent from within carries symbolic weight. It suggests that the ethical questions surrounding AI are no longer hypothetical thought experiments. They are operational concerns debated in real time inside the organisations shaping this technology.</p>
<p>AI represents the sharpest version of that tension yet. These systems are not merely platforms for expression. They are cognitive interfaces, tools that mediate thought, decision-making, and increasingly, emotional reflection.</p>
<p>When the architects themselves begin voicing concern about incentives, misuse, and scientific overreach, it may be worth pausing the applause. Because progress without guardrails is not neutrality. It is acceleration.</p>
<p>And acceleration, without deliberate restraint, has rarely been humanity’s safest instinct.</p>
<p>So, yes, circling back to the sort-of attention-grabbing headline, any internal rebellion will not really help solve anything until companies stop becoming greedy. And this author strongly believes even if pigs could fly one day, profiteering isn’t going to stop. </p>
<p><em><strong>Digital Disconnect</strong> is an <strong>ABP Live</strong>-exclusive column, where we explore the many admirable advancements the world of tech is seeing each day, and how they lead to a certain disconnect among users. Is the modern world an easier place to live in, thanks to tech? Definitely. Does that mean we don’t long for things to go back to the good-ol’ days? Well, look out for our next column to find out.</em></p>
About The Author
