Cognitive warmup. Your Google TV-based Smart TV is set to get more of an AI flavour in 2026. Google, at CES 2026 last week, demonstrated how Gemini on the Smart TV platform would be useful in managing the TV’s settings, searching for specific people or places in photos, and also editing photos into different styles.
Google’s Veo and Nano Banana models will be on hand to manage AI generations too using prompts. Telling Gemini that “the screen is too bright” will be enough to prompt AI to adjust the TVs picture settings (I wonder if there will be a choice to temporarily change settings for that content, or alter the source’s settings completely). Let’s see how this works out, once TV makers begin to roll this out in the coming months.
ALGORITHM
This week, I must slice through three dominant AI narratives currently doing the rounds—that AI is already destroying jobs, AI assistants are unambiguously improving productivity, and that “edgy” AI outputs are an acceptable cost of innovation.
- Oxford Economics shows how AI layoffs are mostly corporate theatre.
- Gmail’s Gemini push may change how you handle emails.
- The Grok controversy—when permissiveness is mistaken for progress.
Taken together, a clear pattern is emerging—that AI today is powerful enough to justify decisions, filter reality, and amplify harm, but not yet mature enough to be trusted with consequences. It is a dangerous place to be.
‘AI layoffs are corporate theatre’
AI is proving a rather easy cover for the real reasons for job cuts.
An analysis by researchers at Oxford Economics punctures one of the most convenient narratives in technology right now—that AI is already wiping out jobs at scale. They say most so-called “AI layoffs” are not driven by AI or automation at all. They are old-fashioned cost-cutting exercises, rebranded with an AI gloss to reassure investors and signal modernisation. If AI was genuinely replacing large numbers of workers, productivity numbers would be exploding. They aren’t. Come to think of it!
What we are seeing instead is symbolic attribution. Companies talk up AI as the reason for workforce reductions because it sounds strategic, inevitable, and doesn’t led to bad publicity. In fact, in corporate speak, it is far better than admitting over-hiring or slowing demand. In practice, AI is still mostly augmentative, inconsistently relevant, and still heavily dependent on human oversight for most tasks.
Distorted policymaking and labour planning, driven by exaggerated claims, may be a greater worry than whether AI can achieve AGI. Here’s something to think about—it seems AI is powerful enough to justify decisions no CEO wants to bear the burden for, but is not yet powerful enough to explain them.
Gmail, a “Gemini era”, and a future trend
Google is firmly of the opinion that Gmail is “entering the Gemini era”. There are two sides to this coin.
First, this is more consequential than just new buttons and summaries in the interface. Gmail is being repositioned from an email service (a communication tool, if you factor in Meets and Chat too) into what seems to be an AI-led summarisation of threads (AI Overviews, something we’ve already seen in Search), suggested replies, help in writing emails (*Help Me Write), realign “important” messages for easier viewing (AI Inbox), and eventually answer questions from within your your inbox.
In other words, Gmail is no longer just managing email—Google is betting on Gemini interpreting intent, relevance, and priority on your behalf.
It all sounds very efficient, but also somewhat unsettling. Email has historically been a system of record. With Gemini in the middle, it may be morphing into a system of inference. If you are happy with letting AI decide what matters, what can be ignored, and how context should be framed, then you’ll be cool with this progression. Google insists this is about reducing cognitive load, but it also quietly normalises delegating judgement to AI.
The bigger issue isn’t accuracy. Once users adapt to AI-filtered inboxes, reverting to raw email chaos may be more difficult than you realise right now, and that’s going to change habits en masse?
Grok, and a warning
This may well be a billionaire’s idea of a joke. It may well be something AI apologists will try to find a silver lining in.
But there is very little that can be considered positive or progressional about xAI’s Grok AI stripping down women (sometimes, even minors) into generated bikinis with a simple prompt by many fairly shambolic collections of blood and organs masquerading as human beings.
What I will add to this is—the Grok controversy isn’t only about a few offensive images, it’s equally about what happens when a platform doesn’t have enough sensibility to distinguish between permissiveness for innovation. Grok’s image generation capabilities were predictably exploited to create sexualised and non-consensual depictions—little surprise regulators and lawmakers in many countries, including India, aren’t at best pleased.
I’ll be honest, I cannot care less about what X, xAI and Elon Musk’s PR machinery’s excuses are. This is a blatant failure on their part. Grok didn’t just generate shockingly problematic images using actual humans as guinea pigs, it also gave the world a live demonstration of how damaging Generative AI systems can get.
Think about this—if your digital devices are caught with child pornography, you’ll face legal action. If X and xAI’s servers host what is essentially generated child pornography, shouldn’t the petulant billionaire go to jail?
THINKING
“The Microsoft 365 Copilot app (formerly Office) lets you create, share, and collaborate all in one place with your favorite apps now including Copilot.”—Banner on Microsoft’s Office.com website
Absolute shambles, in two distinct parts.
First, the hue and cry on social media likely started when Perplexity’s account on X posted that Microsoft has renamed Office to ‘Microsoft 365 Copilot app’ on 5 January — if you still need an example of AI “slop”, this regurgitation may be it.
Second, going by Microsoft’s own wordings which I have mentioned at the very beginning, someone saw it, thought this is new, and of course chose to broadcast it on social media.
But I distinctly remember seeing similar Microsoft 365 Copilot branding on Office mobile apps a few months earlier. Full disclosure: I’m not a Microsoft 365 subscriber and hope to get through this lifetime on that peaceful track.
That said, Microsoft isn’t blameless, and any official statements to the contrary would purely seem akin to damage control. I’ll give you an essence of the disarray.
MS Office was rebranded to Microsoft 365 at some point in 2022, because I’m sure some marketing folks in meetings that probably dragged on for hours would’ve suggested synergies and whatnot as a bundle of apps and services. Then, Microsoft renamed Microsoft 365 to Microsoft 365 Copilot a year ago—if memory serves me well.
I don’t know what the logic was then (perhaps there wasn’t much substance) but to likely make AI into a revenue stream at some point.
There is also a new Copilot chat within Office apps, whether you want it or not. The attempt seems to have been to smartly position the Microsoft 365 paid and free tier subscribers to AI users and pretend all this is very popular. After all, Microsoft 365 Copilot subscriptions stood at ~8 million paid users as of August 2025, according to reports. That’s less than 2% of 440 million Microsoft 365 commercial subscribers.
Note, the Microsoft 365 Premium plan— ₹19,999 per year or ₹1,999 per month—was launched in October, with the highest usage limits for select Copilot features.
The thing is, rebranding an existing suite and forcing AI down everyone’s throats (we’ve talked about this earlier as well, on Neural Dispatch and Wired Wisdom) doesn’t make the Copilot AI popular. In fact, Microsoft’s persistence with forcing Copilot into apps such as Word, Edge and even the Windows 11 OS has proven extremely unpopular.
Truth be told, I am now waiting for the way when the OS will be called Windows 11 Copilot. As things stand, that day isn’t far.