10 Predictions for AI and Ed in 2026

These are in no specific order and I'll try to include a note on accountability - but based on my reading and research, here's where I think we will go in 2026.

The fundamental hallucination error of LLMs won’t be fixed but will be worked on

I increasingly agree with those on the outside that big data and training cannot fix the hallucination problem - a crucial deficit for broader AI adoption and deployment - especially in sensitive industries. In 2026, this will continue to become more broadly recognized; the hallucination error will drop even while still existing and alternate methods will gain credence.

Accountability: At least one major (>$1 billion market cap) AI company* will announce an alternative approach (either beginning construction, planning, development, or deployment) to generative AI apart from the current LLM model.

The egos of those involved won’t die down

AI is the biggest ego race of the modern era and it's only a surprise that Jeff Bezos hasn't swashbuckled his way into it.

Accountability: Altman, Zuckerberg, and Musk will not take a picture all together.

Created via ChatGPT. Not happening.

Federal policy will remain nonexistent, states will continue to step in

The Trump administration will remain under the thrall of David Sacks and hold the line on federal enforcement. States will continue to adopt varying legislation on both the blue and red sides of the divide; the Supreme Court will rule that the EO banning state legislation is unconstitutional. 2027 will be the year AI pushes for legislation. The biggest weakness here will be if the 2026 midterms become a referendum on AI and its economic and societal effects -- then a lame-duck Republican Congress may kowtow the lobby and Trump admin and pass something.

Accountability: No federal legislation in 2026.

Disparity in teacher usage will grow even as PD increases because of a failure of systematic implementation

AI can change teacher and student workflows. Some have jumped on board. Others have resisted. Many await guidance and a push. Many school leaders will continue to lead from behind, nervous about their own capabilities, initiative fatigue, and a lack of systematic clarity.

Accountability: I will offer more PD sessions in 2026 than 2025 on AI, but greater than 90% of those will still require an "Intro to AI" component.

Companion use will grow and parent orgs will start to notice (social media and Australia)

Australia was the kick in the pants many other countries needed and social media companies will be facing increasing limitations around the world. As AI companion use grows - parents and organizations will notice the artificial friendships occupying time and screens that real-life and digital-life ones used to do.

Accountability: There will be legislation in at least 5 US States modelled on the social media laws of Australia. Australia (and similar foreign countries pursuing such legislation) and at least 1 US State will begin legislation to include at least one (>$500 million market cap) AI companion-focused company (or general-use model that won't age-restrict chats).

 Incremental gains will continue and writing/reading will decline

AI mainly got better in 2025, with no major major leaps. We also read and write less. This will continue.

Accountability: This isn't really controversial; but I guess outside measuring sources such as polls and AI metric websites.

Slop limit will be hit but edge cases on reality will be blurrier

Those on social media are tuning out the waves of AI slop that rising more and more. Fortunately, like ads on search results; we will adapt and ignore the most obvious - clicking through for real human connection. However - truly well-made (artisanal AI images or video, if you will) AI will become indistinguishable from human-produced content except by experts.

Accountability: A scandal based on AI-created video or image will erupt in 2026. You will get tired of the AI-produced images in this blog.

There will be more noise. So your marketing and mission will need to focus on cutting through it.

The flipside of the rise in slop will be the increase in noise. We are inundated already, and it will be harder to both make sure your message is publicly available and accurately represented by AI. You need to establish a school-wide communication approach that acknowledges this; and one that teachers, parents, and other stakeholders understand and are bought in to.

Accountability: Fewer responses to social media posts; a decline in participation rates as more and more feel overwhelmed.

The Vatican will release or call for AI related ethics

I expect that Antiqua et nova will be the first in a Theology of the Body-esque compendium from the Vatican over the next twenty years (A Theology of the Mind?!). This will start this year with either a second DDF document building on the conceptional beginning of January's document or a full-blown encyclical from Pope Leo XIV.

Accountability: Pope Leo XIV issues Imago Dei, the first, true encyclical of his reign which presents the nature of man's mind in contrast to that of creation.

Many schools won’t see the big picture of societal/ethical/academic/economic change but focus on one of those in a boomer or doomed way

You cannot hide - AI is threading its tentacles for good and for bad in every way. This will only accelerate but our schools are still stuck and limited and lacking overall coordination.

Accountability: I will offer more PD sessions in 2026 than 2025 on AI, but greater than 90% of those will still require an "Intro to AI" component. Media reports of increased disciplinary issues related to AI glasses and toys will rise noticeably (+500% in google trends) in 2026.

There will be on major black swan event or breakthrough that is AI related in 2026 but it will not be AGI (or ASI).

Talk about ass-covering! But yes - I predict the unpredictable. If 2025 was supposed to be the year of agents (which didn't pan out) but became the year of incremental and noticeable progress; there will come something out of left-field in terms of AI. I'm just not smart enough to predict it yet! Looking back at this year, the closest thing might be the Deep Seek moment of January/February.

Accountability: A major (>$1B) AI-related company* will lose or gain over 50% of its value on the stock market over a 48 hour training period following a (real) news report.

*This seems like a good list of major AI-related companies.