AI Challenges

AI has become a daily news topic—and it is likely to remain one. Both the President[1] and the Pope[2] have emphasized that public awareness and understanding are essential as AI evolves. Current AI capabilities are doubling roughly every four months[3], outpacing even “Moore’s Law” for chip technology. What seems advanced in July may be eclipsed by December, creating new opportunities and new risks each cycle. At least seven impact areas call for broad public awareness and discussion.

1. Jobs

AI will significantly reshape the job landscape. Unlike the industrial revolution—which was constrained by physical resources and information flow—AI applications can be deployed globally with minimal incremental cost once developed. Every organization needs to identify AI opportunities and establish clear usage guidelines. Employees should understand their company’s policies and begin learning how AI can enhance their productivity. Experimenting outside of work and taking relevant courses will provide long-term career resilience. At a societal level, we must consider how to strengthen the safety net for those affected by AI-driven transitions. Education and public investment will be key to developing our workforce and supporting those affected by AI[4]. This is also the focus of a new bipartisan state-action focused organization RAISEUS[5].

2. Misleading Content

Search engines, social media platforms, and other digital tools increasingly incorporate AI into their interactions with us. These systems can summarize answers, shape our information diet, and even act as sympathetic companions—while collecting personal data, addicting users[6] and subtly influencing decisions[7] in ways that may align more with advertisers[8] or funders than with our best interests. We risk drifting down a primrose path and surrendering personal agency. A strong dose of skepticism and critical thinking is essential at a time when the technology often encourages the opposite.

3. Security: Personal, Government and Corporate Devices

A recent AI system—Mythos from Anthropic—reportedly can identify thousands of exploitable flaws in existing software systems[9]. These vulnerabilities span current versions of Windows, Linux, Android, and Apple platforms, as well as older systems. Our phones and computers are at risk if these flaws are exploited by malicious actors. Anthropic is working with vendors[10] so they can issue patches before this occurs. Tests indicate classified US system vulnerabilities are among those identified[11]. In June the US Commerce Department imposed export controls on Anthropic Mythos 5 and Fable 5 resulting in these being removed from the market.[12] Other AI models are not far behind in identifying and exploiting these flaws. Unfortunately, older systems that no longer receive updates likely will remain vulnerable. The key public takeaway: stay aware and apply updates promptly as vendors release them.

4. Software Quality and Ethics

AI systems are increasingly used to generate software. One longstanding weakness in human-developed software is the inconsistent application of engineering standards and best practices. Decades ago, software engineering was a licensed profession—similar to civil or electrical engineering—before being discontinued due to low demand. Poor software engineering has cost lives. These best practices matter just as much for AI-generated code as for human-written code.,

Professional societies such as IEEE[13] and ACM[14] have established ethics guidelines, international standards[15] and certifications[16],[17] that can help ensure software quality. AI could be used to enforce these standards, just as it can detect security flaws. As AI becomes embedded in more aspects of daily life, it would be reassuring to know that the resulting systems are engineered to protect human health and safety.

5. AI Data Centers

AI’s rapid evolution requires massive data centers that consume significant electricity[18] and large volumes of water[19] for cooling. Communities and states must consider not only the immediate impacts but also the likelihood that some facilities may be abandoned as companies compete for survival over the next decade. Upfront protection for infrastructure costs, along with thorough analysis of impacts on electrical and water resources, is essential to prevent communities from bearing externalized costs.

6. Warfare

AI is increasingly used to identify, evaluate, and target advanced weapon systems[20],[21]. Reducing civilian casualties and preventing friendly-fire incidents are clear benefits. But removing humans from the trigger may also reduce moral friction or hesitation, potentially making it easier to initiate or escalate conflict rather than pursue alternative paths to resolution[22].

7. Existential Threat

A 2025 RAND report[23] concluded: “Although we could not show in any of our scenarios that AI could definitely create an extinction threat to humanity, we could not rule out the possibility.” The report suggests that “resources dedicated to mitigating extinction risk are most valuable when they also strengthen defenses against global catastrophic risks and improve AI safety more broadly.” Various corporate leaders and AI experts have voiced related concerns[24]. A sufficiently capable AI system could “decide” that humans are an inefficient use of resources and seek to reduce that impact[25]. Serious attention must be paid to the evolution of AI systems—and we must be able to detect when a system is approaching this level of capability.

A Call to Action

We the people need to be aware and informed about the capabilities and potential risks associated with AI. We must invest time and effort to remain cognizant of the challenges and opportunities that AI represents. We need to influence the decision-making processes that affect us in our workplaces and public institutions. And we need to call upon the educational system to provide appropriate AI literacy and critical-thinking skills at all levels.

Jim Isaak, Londonderry – 30 June 2026
updated & annotated version of 29 June 2026 OpEd in the Manchester Union Leader

[1] https://www.whitehouse.gov/presidential-actions/2026/06/promoting-advanced-artificial-intelligence-innovation-and-security/ retrieved 25 June 2026

[2] https://www.vatican.va/content/leo-xiv/en/encyclicals/documents/20260515-magnifica-humanitas.html retrieved 25 June 2026

[3] https://www.anthropic.com/institute/recursive-self-improvement retrieved 25 June 2026

[4] https://ieeeusa.org/assets/public-policy/positions/ai/AI_and_Electric_Grid_0525.pdf IEEE USA Policy Nov 2024

[5] https://www.raiseus.ai/ retrieved 25 June 2026

[6] https://spectrum.ieee.org/social-media-trial 27 Mar. 2026

[7] https://ieeexplore.ieee.org/document/11371754 P7008/D19 , Jan 2026 - IEEE Draft Standard for Ethically Driven Nudging for Robotic, Intelligent and Autonomous Systems

[8] https://ieeeusa.org/assets/public-policy/positions/communications/SocialMediaRecommendations0224.pdf Feb. 2024

[9] https://www.anthropic.com/news/claude-fable-5-mythos-5 retrieved 25 June 2026

[10] https://www.anthropic.com/glasswing retrieved 25 June 2026

[11] https://apnews.com/article/anthropic-mythos-ai-classified-systems-vulnerabilities-testing-3e8762c0527c4d8ed657cbe48c84a718 retrieved 25 June 2026

[12] https://www.anthropic.com/news/fable-mythos-access retrieved 25 June 2026

[13] https://techethics.ieee.org/resources/ieee-frameworks/ retrieved 25 June 2026

[14] https://www.acm.org/code-of-ethics retrieved 25 June 2026

[15] https://standards.ieee.org/wp-content/uploads/import/documents/tocs/softeng.pdf retrieved 25 June 2026

[16] https://standards.ieee.org/products-programs/icap/ieee-certifaied/professional-certification/ retrieved 25 June 2026

[17] https://www.computer.org/education https://www.computer.org/education/certifications?source=sab retrieved 30 June 2026

[18] https://ieeeusa.org/assets/public-policy/positions/ai/AI_and_Electric_Grid_0525.pdf May 2025

[19] https://harvardsciencereview.org/2026/02/28/re-architecting-the-ai-server-the-hidden-water-cost-of-data-centers-part-ii/ February 28, 2026