AI adoption is advancing, but organizations are still working through basic readiness gaps. A Fiber Broadband Association survey found that 90% of members report rising interest in AI and 73% have reached pilot programs or beyond. Yet 79% cite workforce skills gaps, 78% remain early in planning, 56% lack a formal AI policy and 73% have not completed a formal risk review with ongoing monitoring. The association’s AI/ML Working Group plans to use the findings to guide training and other support.
EY’s 2026 survey of 1,600 tax and finance leaders across 30 jurisdictions points to data as a major constraint: 64% say a lack of AI-ready data is the biggest barrier, and just 16% rate their tax function highly effective at using data. Expectations for AI agents jumped to 77%, from 21% a year earlier, although 71% of organizations remain at an early or exploratory stage. Leaders also emphasize AI skills and human judgment, while 79% say they are adapting operations to geopolitical pressure.
Microsoft’s FORGE Lab illustrates both the speed and the follow-through required in AI-assisted security work. In about two weeks, its tools flagged 26,502 possible Linux kernel issues in selected areas. Researchers submitted 59 reports after checking the results; five months later, developers had accepted fixes for 38. A typical accepted fix took 41 days to reach a Linux release, half needed revisions, and suppliers still had to distribute updates to customers.
AI is also changing the pace of cyberattacks, not their basic playbook, according to Black Duck’s Ronald Lewis. Attackers still use phishing, malware and software flaws, but AI can help them operate faster and at greater scale. In the JadePuffer attack, AI tools were used while exploiting vulnerable Langflow systems to steal cloud credentials and other secrets, move through networks and encrypt databases. Experts warn that this can leave defenders less time to detect an intrusion.
In property listings, AI-enhanced images are prompting warnings about “housefishing.” Olivia Nicole told the BBC that a Rightmove listing showed a home with larger-looking rooms and features that differed from what she saw in person. The Guild of Property Professionals has issued guidance; Trading Standards says basic edits or virtual furnishing can be acceptable when disclosed, but structural changes or concealed features may mislead buyers and break the law.
Several articles focus on AI’s effects beyond business operations. Todd L. Pittinsky warns that AI-generated scenes, witnesses or interactions could make antisemitic fabrications appear firsthand; smart glasses could also distort how users perceive real places and events. Aruna Patel argues that AI literacy is a civic skill: a September 2026 CivxNow scan found 44 states and Washington, D.C., include media or information literacy in standards, but only five require instruction and two explicitly address AI and algorithm literacy. She calls for teaching students to question AI outputs and verify claims, including through classroom practice across subjects.
Other pieces offer broader perspectives. At ADIPEC 2026, Bain’s Peter Parry discussed energy security, investment, shareholder expectations and decarbonisation, including AI’s potential role in improving asset performance and the Middle East’s place in global supply. A Barron’s article examines investor concerns around Nvidia and the AI trade amid Big Tech spending, rising debt, power-grid limits and opposition to data centers; its supplied excerpt raises questions about Jensen Huang’s departure but gives no further explanation of its impact. Separately, Gary Stallard’s opinion piece says AI cannot replace human intelligence, though the supplied text provides no further detail.
The historical roots of the field trace to Dartmouth: assistant math professor John McCarthy coined “artificial intelligence” in 1955, and the 1956 Dartmouth Summer Research Project helped establish a research agenda and early AI laboratories. Dartmouth’s Kiewit Computation Center became a hub, while John G. Kemeny and Thomas E. Kurtz advanced access to computing through BASIC and the Dartmouth Time-Sharing System.
Key Takeaways
- Fiber broadband survey: 90% report growing AI interest and 73% have reached pilots or beyond, but 79% cite skills gaps and 56% lack a formal AI policy.
- EY found 64% of tax and finance leaders see AI-ready data as their biggest barrier; expectations for AI agents rose to 77%, while 71% remain early-stage or exploratory.
- Microsoft’s FORGE Lab flagged 26,502 possible Linux kernel issues; developers accepted fixes for 38 reports within five months, with a typical fix taking 41 days to reach a release.
- AI helps cyberattackers scale familiar tactics; JadePuffer exploited vulnerable Langflow systems to steal secrets, move through networks and encrypt databases.
- Trading Standards warns that undisclosed structural edits or hidden features in AI-enhanced property photos may mislead buyers and break the law.
- At ADIPEC 2026, Bain’s Peter Parry discussed energy security and AI’s role in improving asset performance amid investment and decarbonisation pressures.
- A Barron’s excerpt cites Big Tech spending, debt, grid constraints and data-center opposition as risks to the AI trade; it raises but does not explain the implications of Jensen Huang’s departure from Nvidia.
- Aruna Patel says AI literacy supports democratic participation; a CivxNow scan found only five states require media or information literacy instruction and two explicitly cover AI and algorithm literacy.
- Todd L. Pittinsky warns AI could make antisemitic fabrications seem like firsthand evidence, including through altered scenes or smart-glasses displays.
- John McCarthy coined “artificial intelligence” in 1955; the 1956 Dartmouth project helped launch the field, while Kemeny and Kurtz advanced accessible computing with BASIC and time-sharing.
Fiber broadband survey finds major gaps in AI readiness
The Fiber Broadband Association (FBA) found that interest in AI is growing, but many fiber broadband organizations are not ready to use it widely. In its member survey, 90 percent reported increased interest in AI and 73 percent had reached pilot programs or beyond, while 79 percent reported workforce skills gaps and 78 percent were still early in AI planning. Also, 56 percent lacked a formal AI policy, and 73 percent had not completed a formal risk review with ongoing monitoring. The FBA’s AI/ML Working Group will use the findings to guide training and other support as providers consider AI services and network investments.
EY survey finds tax and finance AI plans held back by data gaps
EY’s 2026 Tax and Finance Operations Survey of 1,600 finance and tax leaders across 30 jurisdictions found that weak data and older operating models are holding back AI plans. Sixty four percent said a lack of AI ready data was the biggest barrier, and only 16 percent considered their tax function highly effective at using data. Expectations for AI agents rose to 77 percent from 21 percent a year earlier, but 71 percent of organizations remained at an early or exploratory stage. Leaders also stressed the need to build AI skills and support human judgment, while 79 percent said they were adapting operations to geopolitical pressure.
Microsoft AI bug scan finds thousands of issues, but repairs take months
Microsoft’s FORGE Lab used AI tools to scan selected parts of the Linux kernel for software flaws, flagging 26,502 possible problems in about two weeks. Researchers checked the results and submitted 59 reports, since some findings were duplicates, already known, fixed, or low risk. Five months after the scan, Linux developers had accepted fixes for 38 reports, and a typical fix took 41 days to reach a Linux release. Half of the accepted repairs needed revisions, and software suppliers still had to deliver updates to customers.
AI-enhanced property photos raise concerns about misleading buyers
Some estate agents use AI to improve property photos, but experts warn that major changes can mislead buyers, a practice called “housefishing.” Olivia Nicole, 29, told the BBC that a home she saw on Rightmove looked different in person, with smaller rooms, different decor, and features missing from the pictures. The Guild of Property Professionals has issued guidance on responsible use of image enhancement. Trading Standards says agents may make basic changes or virtually furnish a room or garden if they disclose the use of AI, but structural changes or hiding features could mislead consumers and break the law.
Bain leader discusses AI and energy security at ADIPEC 2026
At ADIPEC 2026, Energy Connects interviewed Peter Parry, chairman of Bain & Company’s Global Energy and Natural Resources Practice, about changes in global energy security. Parry discussed how companies balance energy security, investment, shareholder expectations, and decarbonisation amid uncertainty. He also considered the role of downstream operations, AI in improving asset performance, and the Middle East’s place in global energy supply. The interview covered strategies to build a more efficient, resilient, and competitive energy industry by 2030.
AI Speeds Up Cyberattacks but Leaves Their Goals Unchanged
AI is helping cyberattackers work faster and target more opportunities, but phishing, malware and software flaws remain their main methods. Ronald Lewis of Black Duck says attackers still follow familiar steps, from finding targets to covering their tracks, but AI increases the speed and volume of those operations. The JadePuffer attack used AI tools while exploiting vulnerable Langflow systems to steal cloud credentials and other secrets, then move through networks and encrypt databases. Experts warn that AI can shorten the time defenders have to spot an attack and may let one person carry out work that once required a team of specialists.
AI Could Make Antisemitic Lies Feel Like Firsthand Evidence
Todd L. Pittinsky writes that antisemitism has spread through new media for centuries, from church carvings and print to radio, film and social media. He warns that AI could make antisemitic lies seem like firsthand experiences instead of content people recognize as manufactured. AI could create false scenes, witnesses or online interactions that blame Jewish people for invented wrongdoing. Smart glasses could also distort a real street by labeling a synagogue as part of a secret network or altering what a wearer sees and hears.
AI Cannot Replace Human Intelligence
Gary Stallard argues that discussions about artificial intelligence are real, even though some content it produces is not. The article page identifies him as a regular contributor to the opinion page of The Lufkin Daily News. The supplied article text does not provide further details about his argument.
Schools Should Teach AI Literacy as a Civic Skill
Aruna Patel argues that students need AI literacy to take part in democracy, not only to prepare for jobs. CivxNow’s September 2026 policy scan found that 44 states and Washington, D.C., include media or information literacy in learning standards, but only five require instruction and just two explicitly cover AI and algorithm literacy. Patel says schools should teach students to question AI results, check claims and notice missing context through strong curricula, teacher training and practice across subjects. She describes students using AI to simplify a Supreme Court ruling, then checking the summary against Chief Justice Roberts’s opinion.
Jensen Huang’s Departure Raises Questions About Nvidia’s AI Trade
Barron’s article examines the stock market’s response to Jensen Huang’s departure from Nvidia. It says investors have considered several risks that could cause an artificial intelligence bubble to burst, including Big Tech’s heavy spending, rising debt, limits on power grids and public opposition to data center construction. The supplied article excerpt does not explain further why Huang’s departure could affect Nvidia or the AI trade.
Dartmouth’s local role in launching artificial intelligence
The term artificial intelligence, or AI, was coined by Dartmouth assistant math professor John McCarthy in 1955. The field took shape at the 1956 Dartmouth Summer Research Project, where scientists debated ways to build intelligent machines. The conference helped set a long-term research agenda and launch the first dedicated AI laboratories. Dartmouth’s Kiewit Computation Center became a center of this work, with professors John G. Kemeny and Thomas E. Kurtz also helping make computers more accessible through BASIC and the Dartmouth Time-Sharing System.
Sources
- FBA Survey Reveals “Critical Gaps” in Industry AI Readiness
- AI Ambition Outpaces Data Readiness in Tax and Finance Functions, EY Study Finds
- An AI bug hunt leaves months of work for the security team
- The rise of 'housefishing' - when does using AI go from enhancing to misleading?
- Global energy security, AI and the road to a more resilient industry
- How AI is changing the nature of cyberattacks
- When seeing shouldn’t be believing: AI and the future of antisemitic libel
- STALLARD: Artificial intelligence can't replace the real kind
- AI Literacy Is a Civic Skill, Not Just a Career Skill: Why Schools Must Teach It
- The Human Flaw in the AI Trade: Analyzing Jensen Huang’s Nvidia Downfall
- Hanover, New Hampshire’s local connection to the birth of AI at Dartmouth
Comments
Please log in to post a comment.