A federal judge in Oklahoma ruled a warrantless Flock license plate search was "indiscriminate mass surveillance" and unconstitutional. OpenAI released GPT-6.1 Sol, which it says nearly matches its most powerful model at a fifth of the price. Anthropic is committing $100 million to train 10,000 engineers who roll out AI inside businesses. And a Tokyo court ruled for the first time that a person's voice can be protected from AI clones.
A federal judge in Oklahoma ruled a warrantless Flock license plate search unconstitutional, calling it indiscriminate mass surveillance. Ashley's take: collect only the customer data you need. OpenAI's GPT-6.1 Sol nearly matches its top model at a fifth of the price. Her take: rerun the numbers on shelved AI projects, but avoid vendor lock-in. Anthropic is investing $100 million to train 10,000 AI deployment engineers. Her take: build your own team's skills now. And a Tokyo court ruled a person's voice can be protected from AI clones. Her take: use licensed voices and protect your own. Listen to the latest AI news you should know about to stay up to date on the latest ai headlines.
Links:
In this episode:
(00:05) Intro: this week's headlines on the Flock ruling, GPT-6.1 Sol, Anthropic's $100M engineer academy, and the Tokyo AI voice case
(00:30) Headlines continued, plus a quick word from our sponsors
(00:46) Let's dive in: federal judge rules a Flock search unconstitutional
(01:11) What an ALPR is, and the Oklahoma ruling
(01:38) How one search revealed a month of someone's travel
(01:56) The ruling and the caveats
(02:18) The Texas Border Patrol case and 100,000 warrantless searches a month
(02:30) Ashley's take: scale changes the rules for your customer data
(03:05) Read your camera-sharing agreements; next up, GPT-6.1 Sol
(03:24) OpenAI's pitch: near-Astra performance for a fifth of the price
(03:50) The price breakdown and where to get it
(04:14) OpenAI's reliability and safety numbers
(04:39) The rest of DevDay: canceled Astra launch and Pro plan changes
(05:00) Ashley's take: rerun the numbers on shelved AI projects (05:24) Avoid lock-in, give AI the least access possible, and a quick break
(06:05) Welcome back: Anthropic launches Claude Frontier Academy
(06:34) Why it exists, and how the residency model works
(06:57) Who it's for and who's in the first cohorts
(07:23) Anthropic's partner network, and Ashley's take
(07:44) Don't wait: invest in your own team's skills
(08:24) Who Kenjiro Tsuda is and what happened
(08:46) TikTok's defense and the court's ruling
(09:06) What publicity rights are, and the twist
(09:33) Why it matters, and Ashley's take: your voice is part of your brand
(09:58) Use licensed voices, and protect your own
Follow Ashley at @ashleyrcoffey89 on Instagram and Threads. Want the highlights without the twice-a-week commitment? Ashley's monthly AI newsletter is linked in her bio. Send us a topic you want covered, and follow the show for AI news you should know about. Thanks for listening!
Ashley Coffey (00:05)
welcome
Ashley Coffey (00:05)
back to AI News that you should know about for October 6th. I'm your host, Ashley Coffey. There's a lot of AI news out there, but we give you the weekly highlights quickly. This week on the show, we'll be talking about a federal judge rules a flock search was indiscriminate, mass surveillance, and unconstitutional. OpenAI releases ChatGPT 6.1 Sol at a fifth of GPT-6 Astros token prices.
Ashley Coffey (00:30)
Anthropic to invest a hundred million dollars to train AI engineer talent, and Tokyo Court grants legal protection to human voices in AI clone case. We'll be covering all of that on our show today, but first here's a quick word from our sponsors.
Ashley Coffey (00:46)
Welcome back. Let's dive in. All right, a federal judge rules a Flock search was indiscriminate mass surveillance and unconstitutional. This is coming from 404 Media and it hits close to home because this comes out of Oklahoma where I live. Let's get into it. Flock runs a nationwide network of automated license plate reader cameras called ALPRs. They log every car that passes and let police search where a vehicle has been.
Ashley Coffey (01:11)
Quick definition: an ALPR is a camera that automatically reads and records license plates. This one's close to home. A federal judge in Oklahoma, Sarah Hill, ruled Thursday that a Tulsa County deputy violated a driver's Fourth Amendment rights when he searched her plate in flock without a warrant. What happened? In May of 2026, the deputy saw an SUV with California plate on an Oklahoma highway and followed it for no apparent reason other than the out-of-state plate.
Ashley Coffey (01:38)
He ran the plate in flock, pulled the driver over for changing lanes without a signal, then compared her answers about her trip to the flock data. He used that travel history as part of his reason to search her car. How much one search revealed? More than fifty records of her whereabouts across the country over an entire month.
Ashley Coffey (01:56)
The ruling. The judge called the network a type of indiscriminate mass surveillance and said tracking someone's whole pattern of movement is a search that requires a warrant. All the flock evidence and everything found in the car gets thrown out. The caveats: the ruling isn't binding precedent and other courts have gone the other way, but it's one of the first times a federal judge has found a flock search unconstitutional.
Ashley Coffey (02:18)
And it comes the same week a jury found a Border Patrol license plate scheme unconstitutional in Texas. The scale, according to audit logs 404 Media Reviewed, police run more than 100,000 warrantless flock searches every month. My take here: this is the clearest example yet of what AI does to data you didn't think was sensitive.
Ashley Coffey (02:40)
One licensed plate photo is harmless. Millions of them stitched together and searchable in seconds becomes a map of someone's life. The judge's point is that scale changes the rules. The same logic applies to the data you collect if you're a business. Customer visits, location check-ins, security food footage, once it's searchable with AI, it becomes far more revealing than what it than when it's sat in the folder. Collect only what you need, decide how long you keep it.
Ashley Coffey (03:05)
and know who has access. And if you or your business or property shares camera footage with police or a vendor network, read the agreement, know whether your camera feeds a bigger system and who can search it. Next up, OpenAI releases GPT 6.1 at a fifth of GPT-6 Astra token prices. Let's talk about it.
Ashley Coffey (03:24)
Callback to two episodes ago, where we covered GPT-6 Sol and Luna landing 90 minutes after Claude Opus 5.5. At its dev day event this week, OpenAI upgraded Sol again with GPT-6 Sol. OpenAI's pitch. Near Astra performance for a fifth of the price. OpenAI says it's nearly matching GPT-6 Astra, its most powerful model on coding, computer use, and professional work. The price.
Ashley Coffey (03:50)
$2 per million input tokens and $10 per million output tokens. Cache input, which is reused context, drops to 10 cents per million. Remember, tokens are the chunks of text AI models bill by, roughly three-quarters of a word each. Where you get it, ChatGPT work and codex for plus pro, business, enterprise, and EDU users, plus the API.
Ashley Coffey (04:14)
It's not in the main GPT chat yet, but OpenAI's own numbers on reliability on deliberately difficult prompts, answers with a factual error dropped from 11.4% to 7.7%. In a safety test, it tried to get around explicit restrictions like access denied messages, and 23.5% of the time, down from 64.4% of the time for GPT-6 sol
Ashley Coffey (04:39)
These are OpenAI's own tests, mostly in low-stake situations. Remember, the rest of the dev day, OpenAI canceled the October launch of GPT-6.1 Astra after a failed safety test, launched a new $500 pro plan, and cut usage on its $200 pro plan in half. My take here, the price war we keep talking about is not slowing down.
Ashley Coffey (05:00)
Near top-tier performance at a fifth of the price is exactly what makes AI automations pencil out for a small business. If you shelved a project because the cost per task was too high, run the numbers again. But you read the fine print on both sides of the announcement. The same week prices dropped in the API, OpenAI halved usage on one of its consumer planes and added a $500 tier. Prices moved in both directions.
Ashley Coffey (05:24)
That's the build so you can switch lesson from the soft bank story from last week. Don't lock into a long contract or single vendor. And tries to get around restrictions 23.5% of the time as an improvement, not a green light. That's the same lesson from the Australia story. Give the AI tools the least access possible. All right, we're going to take a quick break, and when we come back, we'll be talking about.
Ashley Coffey (05:47)
Anthropic to invest a hundred million dollars to train AI engineer talent, and Tokyo courts grant legal protection to human voices in AI clone case.
Ashley Coffey (06:05)
Welcome back. Continuing with our stories. Anthropic to invest $100 million to train AI engineer talent. Let's talk about it. What it is, Anthropic launched Claude Frontier Academy backed by a $100 million commitment to train 10,000 Frontier deployed engineers by the end of 2027. Quick definition: a deployed engineer is someone who works inside a business to actually build and roll out AI, not just recommend it. That's exactly what I do in my business with Coffey and Consult. This is it.
Ashley Coffey (06:34)
Why? Anthropic says the biggest roadblock to business getting value from AI right now isn't the models, it's the people who know how to put them to work. How it works? It's modeled on medical residency. Engineers learn from practitioners, practice on realistic cases, and pass an assessment before they earn the credential. It starts with a multi-day in-person intensive with anthropic engineers, including a simulated enterprise deployment that gets graded.
Ashley Coffey (06:57)
Who is it for? Hands-on software engineers with strong fundamentals, experience with building AI models, and experience helping others adopt AI. Prior experience building AI agents isn't required. Who's in the first cohorts? Engineers from Accenture, Bain, Cap Gemini, Deloitte, Mackenzie, Morgan Stanley, Novo Nordisk, and Commonwealth Bank of Australia, among others. Cohorts are running in San Francisco, New York, and London.
Ashley Coffey (07:23)
What it builds on, Anthropic's partner network where professionals across 46,000 firms have already earned more than 175,000 Claude certificates. My take here, Anthropic is saying out loud what I see with clients every week. The technology is ready, but the people and processes usually aren't. Buying the tool is the easy part, but getting into real workflows is the hard part.
Ashley Coffey (07:44)
notice who's in the first cohorts. Big consulting firms and enterprises. Those 10,000 engineers will mostly serve big companies first. For a small business, the takeaway isn't to wait for one of them, it's to invest in your own team skills now, because the businesses that build the muscle internally will move faster and spend less.
Ashley Coffey (08:02)
And when you hire outside help, ask what training and credentials they actually have. "We use AI" asn't a qualification anymore. Also worth noting, this is good for the industry, and it's also good for Anthropic since every engineer trained here builds on Claude. All right, next up, the Tokyo courts grant legal protection to human voices in AI clone case. This is coming from the Associated Press. Let's talk about it.
Ashley Coffey (08:24)
Who's who? Kanjiro Suda is a famous Japanese anime voice actor known for jujutsu, kaizen, and Yu Gi What happened? An anonymous TikTok account posted at least 180 videos about urban legends and conspiracy theories narrated by a deep voice Suda said was an AI clone of his. He sued TikTok's operator in November 2025 to get the videos taken down.
Ashley Coffey (08:46)
TikTok's defense, it called the narration a generic male voice and said any resemblance was subjective. It also pointed to the uploader's claim that the AI was trained on a friend's voice. The ruling on Wednesday, the Tokyo District Court ruled that a person's voice can be protected under publicity rights, Japan's first ruling of its kind.
Ashley Coffey (09:06)
The judge said that protection applies when a voice is used purely to cash in on its commercial appeal, comparing a voice to a portrait. Quick definition publicity rights protect a person's control over commercial use of their identity, like their name or their face. The twist here, Suda didn't get his takedown because the account was already deleted in May while the case was underway, and the court didn't decide whether the voice actually copied his. Why this matters though?
Ashley Coffey (09:33)
Japanese voice actors now have a precedent to cite against AI clones, and the industry has launched a no more campaign against AI imitation without permission. Suda says the case achieved what it set out to do and he won't appeal. My take here: your voice is part of your brand, and courts are starting to treat it that way. That matters if your business uses AI voices anywhere, including ads, social videos, phone menus, or the AI phone agents we've been talking about on this show.
Ashley Coffey (09:58)
The rule for your business: use licensed AI voices with written permission and never ask for an AI voice that sounds like a celebrity or anyone else. The legal risk is real and so is the trust risk if customers figure it out. And if you're the face and voice of your business on a podcast and videos on sales calls, your voice is an asset worth protecting too.
Ashley Coffey (10:17)
Know where your recordings live and read the terms on any AI tool that you upload them into. That's it for AI News That You Should Know About This Week. Show notes have everything that we have covered. And if you want more of this, that's what social media is for. You can find me on Instagram and threads at AshleyRCoffey89.. Follow the show, send us a topic you want covered, and we'll see you next week for more AI news that you should know about. Thank you for listening.