- Everyday AI
- Posts
- Ep 692: The Rise of AI Assistants: The undeniable upside and the ugly downside
Ep 692: The Rise of AI Assistants: The undeniable upside and the ugly downside
The ugly downside of AI agent, OpenAI and Anthropic closer to IPO, Microsoft Moves Toward $500M Anthropic Investment, GPT-5.2 builds its own browser and more
Everyday AI
January 15, 2026
👉 Subscribe Here | 🗣 Hire Us To Speak | 🤝 Partner with Us | 🤖 Grow with GenAI
Outsmart The Future
Today in Everyday AI
8 minute read
🎙 Daily Podcast Episode: AI assistants are moving from helpful tools to always-on workers. The gains are massive, but so are the risks if companies aren’t paying attention. Give it a watch/read/listen.
🕵️♂️ Fresh Finds: OpenAI Releases GPT-5.2 Codex, ChatGPT, Loses WhatsApp Access, OpenAI Pulls Ahead in Enterprise AI Spend and more Read on for Fresh Finds.
🗞 Byte Sized Daily AI News: OpenAI and Anthropic closer to IPO, Microsoft Moves Toward $500M Anthropic Investment, and more Read on for Byte Sized News.
💪 Leverage AI: AI agents won’t just help customers—they’ll stress-test support systems at scale. The downside is arriving faster than most teams expect. Keep reading for that!
↩️ Don’t miss out: Did you miss our last newsletter? We Covered: Google’s big AI move with Personal Intelligence, OpenAI and Cerebras strike $10 billion deal, Google open sources healthcare models and more. Check it here!
Ep 692: The Rise of AI Assistants: The undeniable upside and the ugly downside
AI agents can do our work. OK, sweet. But they can also do.... a lot of bad. Yikes. 😳
So how should enterprises find the middles.
Cuz let's be honest. If your company is not using AI agents in 2026, you're going to fall FAR behind.
But if you don't have the 101 of security and guardrails on point, those AI agents can do more harm than good.
Wish you could get the answers directly from a Fortune 500 CEO?
That's exactly what Concentrix's Chris Caldwell is going to provide. Make sure to join us!
Also on the pod today:
• Deepfakes attacking customer support 🎭
• Age divides in AI trust 👴🧑
• Human-only brands: future trend? ✋
It’ll be worth your 28 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Here’s our favorite AI finds from across the web:
New AI Tool Spotlight – TeamOut Has you Briefly describe your event and it will find the perfect venue in seconds, folk Assistants is Assistants working for you 24/7. Even when you sleep, Notto is Your Invisible AI Assistant for Work
AI Space Training Boost — Slingshot Aerospace scored $27 million to bring AI-driven space conflict training to the Space Force. See how TALOS changes the game.
GPT-5.2 Codex Released — OpenAI just dropped GPT-5.2 Codex for developers, promising smarter coding and top-tier cybersecurity help. Curious about the price jump and new features?
AI Movies — Matthew McConaughey just trademarked “alright, alright, alright” to crack down on AI using his famous catchphrase without permission. Curious how this could change celebrity voice protection?
NotebookLM Tables — NotebookLM tables hint at a smarter, more customizable workflow.
AI In Sports — NBC Sports is adding AI-powered athlete tracking to live games. Find out how it could change your sports streaming.
ChatGPT Taken Off Whatsapp — ChatGPT shutters WhatsApp access after Meta’s sweeping AI ban
Gemini Search Trend Tool — Google Trends now uses Gemini to surface smarter, faster insights. Curious how your searches stack up?
Grok Controversy Update — Grok’s controversial AI tool for undressing images is getting blocked on X after public backlash. See why Musk decided to pull the plug.
Monica 9.0 Launch — Monica 9.0 is here with deep research, slick slides, and browser automation. Want to see what makes it different?
OpenAI Business Surge — OpenAI recordly surges ahead in business AI spending. Curious why ChatGPT is winning over companies again?
1. Microsoft’s Big AI Bet On Anthropic on Track to $500M Mark 🤑
Microsoft is reportedly on track to spend $500 million on Anthropic AI, highlighting the tech giant’s aggressive push to stay competitive in the artificial intelligence race.
This major investment reflects the high stakes and intense competition among top tech companies to secure partnerships with leading AI startups. As Microsoft doubles down on its AI strategy, the move signals just how crucial advanced AI has become to its future plans.
2. OpenAI and Anthropic IPO Buzz Heats Up 🚀
The race for blockbuster tech IPOs is on, as OpenAI, Anthropic, and SpaceX all move closer to going public, according to The New York Times. These three giants are poised to make history with valuations that could rival the biggest listings ever seen on Wall Street.
If they go through, 2026 could be remembered as the year Silicon Valley truly took over the stock market spotlight. With artificial intelligence and space tech leading the charge, investors are bracing for what some are already calling an IPO supercycle.
3. GPT-5.2 Builds Browser in a Week, Shakes Up Coding World 🧑💻
In a stunning leap for AI-assisted programming, Cursor’s team used GPT-5.2 to build a fully functional web browser from scratch in just seven days, skipping traditional browser engines entirely.
The AI generated over 3 million lines of code in an uninterrupted marathon, showing just how rapidly coding automation is advancing. While the browser is still rough around the edges and trails far behind giants like Chrome, the project signals that AI agents are starting to tackle software once thought far out of reach.
4. AI Slip-Up Fuels UK Police Scandal ⚠️
A top UK police official admitted this week that Microsoft's Copilot AI generated fake soccer match details, leading to Maccabi Tel Aviv fans being barred from a major Europa League game in November.
The blunder, which the West Midlands Police failed to catch, sparked political fallout and public outrage, with calls for accountability reaching the Home Secretary. Microsoft's Copilot had apparently hallucinated a match that never happened, but its incorrect data influenced a controversial security decision.
5. AI Godfather Bengio Unveils New Safety Approach, Boosts Optimism 📈
Yoshua Bengio, a key architect of modern artificial intelligence, says his team has developed a framework for making AI systems safer and more controllable, tackling worries about unpredictable behavior and misuse.
Speaking to reporters, Bengio explained that these new methods could help prevent AI from causing harm, and have made him much more optimistic about the long-term impact of the technology. The announcement comes as concerns about AI safety reach new highs, with governments and tech giants scrambling for solutions.
6. OpenAI Reclaims Two Key AI Innovators 🗝️
OpenAI has just lured back Barret Zoph and Luke Metz, cofounders of Thinking Machines Lab, in a shakeup that’s causing waves across Silicon Valley.
The pair’s abrupt exit is seen as a setback for the ambitious new lab, with two competing stories already swirling about their reasons for jumping ship. Their return signals OpenAI’s continued magnetism for top talent at a moment when the AI talent war is heating up.
Your entire customer experience strategy assumes one angry human can only send one complaint at a time.
Wrong.
Agentic AI changes the math of customer service from conversation to asymmetric warfare. We aren't talking about chatbots answering FAQs anymore. Nah. We’re talking about a single disgruntled user commanding a personal bot swarm to hammer your support channels every 15 minutes until they break your queue.
Yikes..
Chris is the President and CEO of global juggernaut Concentrix and he pulled up to today’s Everyday AI show to expose the "ugly downside" of the agentic future.
Because you can pat yourselves on the back for deflecting human calls with an AI agent, but you might be completely unaware of the ugly downside of those same AI agents silently going rogue.
We ripped the lid off this security nightmare on today's show and Chris shared the secrets.
Business leaders need to get ready for a world where thousands of AI bots are chatting with your brand all at once.
Here’s your gameplan, shorties.
1. Bot Swarms Are Coming For You 🚀
You think your call center volume is high now?
Wait until mid-2026.
Chris painted a picture of the near future where a single consumer uses agentic AI to apply for mortgages at 100 banks simultaneously. That ain't just a lead. That is a straight-up denial of service attack on your underwriting team if you aren't ready.
But the real threat is the weaponization of petty complaints.
Customers will soon have personal agents capable of harassing your brand across every channel. Voice. Email. Social DMs. They won't stop until they get what they want. It is an asymmetric battle where one angry guy with an eight-dollar grievance can generate the workload of a thousand tickets.
Most businesses are building AI to answer human questions.
Dead end.
You need to be building "microservice" layers where your AI talks directly to their AI without ever involving natural language. If you force a bot to talk to a human agent then your operational costs will explode overnight.
Try This
Audit your customer service friction points this week with a "Bot Attack" mindset.
Ask your CTO this specific question: "If 5,000 users automated their ticket submission process tomorrow would our system crash or queue them?"
If the answer is queue then you are in trouble.
Start looking into API layers that allow verified consumer agents to resolve low-level disputes via data exchange rather than voice or chat.
Build the fast lane for bots now or they will clog the slow lane for your humans later.
2. Low Tech Fixes High Tech Fraud ⚡
The days of verifying identity with "mother's maiden name" are dead.
With voice cloning technology available to anyone for free a bad actor can sound exactly like your best customer in five seconds. We aren't talking about sophisticated hackers anymore. We are talking about basic scams running on autopilot.
Companies are rushing back to "dial-back" protocols where support hangs up and calls a secure number to verify identity. It sounds archaic but in a world of deep fakes your fancy biometric voice analysis might just be an open door for fraud.
Sus.
If your security relies on static data points or voice recognition alone you are wide open to agentic impersonation. The smartest companies are moving toward dynamic multi-factor authentication that requires physical device confirmation instead of knowledge checks.
Try This
Test your own security team's vulnerability to voice cloning.
Have a co-worker clone your voice on ElevenLabs (it takes five seconds) and try to reset a password or access information from your internal help desk.
When they inevitably succeed use that failure to secure budget for physical token verification or dial-back protocols.
Update your customer verification scripts immediately to remove static questions like "high school mascot" or "street you grew up on."
Those are public data points that AI agents already scraped years ago.
3. Stop Summarizing And Start Doing 🔥
Here is the most brutal insight from the conversation.
Most executives making decisions about AI strategy don't actually use the tools.
They have assistants. They have teams. They view AI as something to be implemented "down there" in operations while they stick to their old workflows. Chris called out the single biggest waste of AI potential he sees in the C-suite which is using it to summarize emails.
If you are using a frontier model just to make your inbox shorter you are missing the point entirely.
The real power ain't summarization. It is action.
Chris uses these tools for complex research and time management. He doesn't use them to shorten paragraphs. If you aren't getting "wow moments" personally you cannot possibly direct a company-wide transformation.
Real talk here.
You can't outsource your brain. If your leadership team isn't hands-on with the tech your strategy will be theoretical, safe, and ultimately ineffective against competitors who live in the tools.
Try This
Stop asking your AI to summarize emails for one week.
Instead you gotta force yourself to use it for "Agency Transfer" tasks.
Give it a goal: "Research the top three competitors to our new product line and draft a comparison table of their pricing models."
Watch where it fails. Watch where it hallucinates.
You need to feel the friction points yourself so you know what your employees are dealing with. If you can't drive the car stop telling the mechanic how to fix the engine.
Reply