🔍 Read the full analysis: Where AI Meets Work: November 14, 2026 on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
A ThorstenMeyerAI.com roundup published Nov. 15 reviews developments reported on Nov. 14, including rising concern about AI among Federal Reserve market contacts, a planned Australian banking stress test and debates over robot models. It also covers stolen Uber driver license data linked to an AI vendor and an FAA outage attributed to contractor maintenance, not AI; several details remain unresolved.
ThorstenMeyerAI.com published a roundup on Nov. 15 connecting recent AI and technology news to work, wages, skills and companies. The report points to a Federal Reserve survey in which 87% of 23 market contacts named AI as a risk, an Australian regulator’s planned AI-focused stress test, and data theft affecting Uber drivers through a vendor. It also distinguishes an FAA outage caused by routine contractor work, according to officials, from AI-related incidents.
The Federal Reserve’s latest Financial Stability Report says 87% of 23 market contacts surveyed named AI as a risk, compared with 50% in the spring. The Fed stressed that these were the contacts’ views, not the central bank’s own assessment. Separately, the report says debt issuance tied to advanced computing exceeded $680 billion by October 2026. Australia’s banking regulator, APRA, plans to examine AI-driven disruption in its 2027 stress test. It has also found that few institutions had tested how they would switch away from a critical AI provider.
The roundup also describes a split in approaches to robotics. Fortune reported that OpenAI had broken up its stand-alone robotics team, with the company saying its general models, including GPT-6 Astra, could be applied directly to robot control. Runway released the specialist robot model Praxis-1 with open weights two days earlier. In Runway’s own test, GPT-6 Astra took 15.6 times longer and cost 51 times more per completed task than Praxis-1. Those figures come from a vendor’s test, not an independent comparison. Separately, Artificial Analysis reported that GPT-6 Astra averaged 30 minutes and $16.36 per task controlling a real robot arm in its AA-Robotics test; a person took about 20 seconds.
Uber has begun notifying US drivers whose data was taken in an October break-in at Fireworks AI, an Uber vendor, according to Uber’s account to TechCrunch. Breach notices reviewed by TechCrunch said driver’s license information was stolen. Notices from three states identified 7,870 affected people, a count the roundup describes as a floor rather than a complete total. The FAA incident was a separate infrastructure story: on Nov. 13, departures at US airports were held for about four hours and 23 minutes. Transportation Secretary Sean Duffy traced the incident to contractor L3Harris switching lines from copper wire to fiber at a remote FAA site, according to news reports. The cause of the resulting problem has not been determined.
AI Risk Reaches Firms and Workers
The developments matter because AI adoption is tied not only to model performance but also to financing, suppliers and operational work. The Fed survey records concerns among market contacts, while the reported debt issuance gives a measure of the capital involved in advanced computing. Neither figure, by itself, establishes that AI is causing financial instability or job losses.
APRA’s planned test makes a practical dependency visible: institutions may need people and procedures to keep operating if a critical AI provider becomes unavailable. For workers, the Uber breach shows how data exposure can travel through a vendor chain. Drivers may have little visibility into which contractors handle their information, while license details can be harder to replace than a password.
Robotics raises a different workplace question. The reported tests suggest a possible trade-off between general-purpose reasoning and speed or cost in a particular task, but the results are not a general measure of performance across workplaces. Deploying robots also depends on real-world operating knowledge and data. The roundup attributes to MIT’s Russ Tedrake the view that robot companies have workplace deployment experience and data that AI labs lack. Workers who operate existing systems may hold some of that practical knowledge.
Popular right now

Hisense 43-Inch E6 Cinema Series Hi-QLED 4K UHD Smart Fire TV (43E6QF)

KitchenAid Artisan Series 5-Quart Tilt-Head Stand Mixer, Juniper (KSM150PS)

eufy C30 Robot Vacuum and Mop
As an affiliate, we earn on qualifying purchases.
The Vendor and Maintenance Chain
The roundup’s central connection is that technology risks often involve more than the AI model itself. Financial exposure can include debt and investment; service continuity can depend on whether a company can replace a provider; and personal information can pass through contractors. These are related concerns, but the events described are distinct and do not establish a single cause or trend.
The FAA outage offers a contrast. Officials linked its start to routine telecom maintenance by L3Harris, not to AI. FAA Administrator Bryan Bedford told CNN there was “zero evidence of anything AI-related.” The FAA’s tracker showed 68% of its telecom connections replaced—3,508 of 5,170—as of Oct. 15, 2026. That modernization work provides background for the incident, but the available account does not identify whether a faulty router or an installation error caused the disruption.
““zero evidence of anything AI-related.””
— FAA Administrator Bryan Bedford, as reported by CNN
The Causes and Counts Still Open
Several limits remain. APRA has not published the numerical assumptions or scenarios it will use in its 2027 stress test. The Fed survey covers 23 market contacts, and its result describes those respondents’ views rather than a universal assessment of AI risk.
The full number of Uber drivers affected is not known from the state notices cited in the roundup; the 7,870 figure is a minimum based on three states. The roundup does not establish the full scope of the stolen information or explain all steps Uber and Fireworks AI have taken in response.
The precise FAA failure also remains undetermined: reports had not established whether a router was faulty or improperly installed. Robotics comparisons cited here come from vendor or test-provider reports and do not settle which approach will work best across real jobs. No source in the roundup establishes the resulting effects on employment, wages or adoption at scale.
Tests, Notices and FAA Findings
The next stated regulatory milestone is APRA’s 2027 stress test, whose scenarios and numerical assumptions have not yet been released. Uber’s notifications to drivers are underway, but the cited material does not give a final affected-person count. Further reporting or company disclosures may clarify the breach’s scope and response.
For the FAA, the outstanding issue is identifying the technical cause of the outage and any corrective measures following the contractor’s line-switching work. Robotics comparisons may also develop as more tests measure speed, cost and reliability in real settings. Until those details arrive, the roundup’s figures are useful as reported snapshots, not definitive predictions about the future of work.
Key Questions
What does the 87% AI-risk figure mean?
It means 87% of 23 market contacts surveyed for the Federal Reserve’s Financial Stability Report named AI as a risk. The Fed said the responses were the contacts’ views, not its own conclusion.
How many Uber drivers were reported affected?
Notices from three states added up to 7,870 people, according to the roundup’s count. That is a partial count, not a confirmed total; Uber’s vendor Fireworks AI was identified as the source of the October break-in.
Was the FAA outage caused by AI?
Officials and news reports linked the outage’s start to contractor maintenance switching lines from copper to fiber at a remote FAA site. FAA Administrator Bryan Bedford said there was “zero evidence of anything AI-related.” The exact technical failure was still unknown.
Do the robot-model tests show which system is better?
No. Runway’s comparison favored its Praxis-1 model on time and cost per task, but it was a vendor test. A separate Artificial Analysis test reported different task measurements. The results do not establish which system performs best across workplaces or tasks.
Source: ThorstenMeyerAI.com
