September 22, 2026 | Ten stories from September 21–22
Today’s AI news ranges from football graphics and movie screens to hospital paperwork and fighter cockpits. The technology keeps finding new places to work. The debate over who should supervise it appears to be hiring, too.
Here are ten stories worth catching up on, with links to the original reporting and announcements. For more throughout the day, visit thedrone.report, my AI news aggregator updated every half hour. The news cycle apparently has unlimited battery life.
1. Twenty countries want stronger human control over AI
A declaration launched by Finland and Norway on September 21 calls for independent evaluations of advanced AI, shared reporting of serious incidents, and exploration of an international institution that could establish standards and support verification. Finland’s presidential office says leaders from 20 countries have joined the appeal, alongside the European Commission’s president. The United States and China are absent from the published list.
This is a call for action, not a newly established global regulator. That distinction matters when the headline starts sounding like someone finally hired a manager for Skynet.
My take: Independent testing sounds useful. Getting competing governments to agree on access, authority, and enforcement is the much harder assignment.
Read the declaration: Office of the President of Finland
2. Bills fans question an AI-generated Prime Video promotion
Buffalo State’s student newspaper, The Record, reports criticism of an AI-generated Bills promotional image shared September 17 in collaboration with Prime Video and other accounts. Its September 22 coverage describes a scene featuring Josh Allen and teammates surrounded by wings, team decorations, and stadium imagery. The reporting includes objections about choosing AI when Buffalo has local artists who could contribute.
My take: There is room for AI artwork and room to question a major brand’s creative choices. The interesting issue is what a team wants its art to communicate about its relationship with fans. Bills Mafia was never likely to treat a marketing graphic as a quiet, private matter.
Read the reporting: The Record
3. Tom Cruise thinks audiences still want the human element
At a BAFTA event in London on Monday, Tom Cruise acknowledged AI’s arrival while arguing that audiences still value human creativity and tangible filmmaking. Variety’s coverage highlights his encouragement for artists to continue creating and his enthusiasm for the shared theatrical experience.
My take: Ethan Hunt has already had an unpleasant professional relationship with an AI villain. His position is understandable. More seriously, a convincing image and a compelling movie are different achievements. Tools can change while audiences continue caring about performance, intention, and craft.
Read the story: Variety Australia, September 22
4. Heidi announces $340 million to expand healthcare AI
Healthcare AI company Heidi announced new financing totaling US$340 million. SiliconANGLE reports that the package combines a $100 million Series C round led by Blackbird with $240 million in growth investment from General Catalyst. The company wants to expand adoption across health systems and move beyond documentation toward completing work under clinicians’ supervision.
My take: Reducing administrative work is an appealing use of AI. Funding demonstrates investor interest, though it does not establish clinical effectiveness or safety. The useful test is whether clinicians gain time without inheriting another pile of errors to correct.
Read the story: SiliconANGLE, September 22
5. Handshake introduces practical AI training for job seekers
The New York Stock Exchange’s September 22 update highlights Handshake’s new AI Skills Studio. It describes personalized, project-based missions that let job seekers work with AI tools used by practitioners. The announcement emphasizes practical skills rather than simply familiarity with AI terminology.
My take: Showing how you checked an AI answer, improved a workflow, or completed a project is more persuasive than adding “AI enthusiast” to a résumé. Enthusiasm is lovely. A working example gives the interviewer something to discuss besides your impressive collection of browser tabs.
Read the announcement: NYSE via PR Newswire
6. Dassault tests new AI aboard the Rafale
Dassault Aviation announced September 22 that it flight-tested two AI algorithms aboard its Rafale fighter. One was developed internally and the other with Thales/cortAIx. The company describes the work as supervised cockpit assistance serving the human crew and says the functions are mature enough for consideration in future upgrades. It does not disclose their detailed tasks in the announcement.
My take: That limited description is worth respecting. It does not establish that AI is independently flying combat missions. We can save the full Top Gun meets Terminator treatment until the actual capabilities justify the trailer.
Read the announcement: Dassault Aviation
7. Adobe and Jet2 want AI to personalize your vacation
Adobe and Jet2 announced a multiyear partnership on September 22. Planned uses include personalized destination suggestions, offers, and itineraries, supported by Adobe’s AI tools. A joint customer-experience lab will develop additional services. Jet2 also plans to use Adobe technology to produce content consistent with its brand.
My take: Useful recommendations could make trip planning easier. Success will depend on whether the system understands what travelers actually need. Recommending a peaceful getaway and then sending someone to a hotel beside a nightclub would be very efficient disappointment.
8. Microsoft expands its AI infrastructure in India
Microsoft’s September 21 announcement identifies its now-live Hyderabad cloud region as a strategic hub for Asia and the Global South. The company says the region has three availability zones, with additional Azure and AI services rolling out over time. Microsoft also reports having provided AI training to more than 10 million people in India toward a 20 million target by 2030.
My take: Infrastructure and training deserve attention alongside model launches. Organizations need reliable access and people who know how to use it. A data center opening is less entertaining than a chatbot meltdown, but potentially more consequential.
Read the announcement: Microsoft
9. Accenture and AWS target midsize businesses
Accenture Edge and Amazon Web Services announced six offerings for mid-market organizations on September 22. Accenture defines the target group as businesses with annual revenue between $300 million and $3 billion. The collaboration combines cloud, security, and AI services with implementation support intended to modernize business operations.
My take: “Midsize” is doing some heavy lifting at $3 billion in revenue. Still, the practical issue is familiar at any size: does the technology improve a process enough to justify its cost? A successful pilot needs to survive contact with the actual business.
Read the announcement: Accenture
10. Survey reporting points to the messy processes behind AI problems
Help Net Security’s September 21 coverage of Sapio Research reports that 40% of the large companies represented in a survey had experienced an AI-related compliance or governance issue during the previous year. The survey involved 1,000 senior leaders, who attributed a role to process problems in 84% of incidents. These are survey findings, not a measured failure rate for every company using AI.
My take: Adding automation to a confusing approval chain can help the confusion travel faster. The unglamorous work matters: deciding who approves actions, recording decisions, and making sure someone can explain what happened afterward. Even Office Space would benefit from an audit trail.
Read the reporting: Help Net Security
Follow the developing stories at thedrone.report. News summaries are attributed to the linked sources. “My take” paragraphs are commentary. Company announcements describe their own products, plans, and claims.
Leave a Reply