AI Homework Help, Tutor Quality and Where Canadians Learn

An independent test of ChatGPT's new Teen Mode found a button that hands a student the answer in one click, inside the very feature built to make them work it out (Education Week, 2026). For parents who have watched homework move onto a laptop, it is a timely reminder that the tool matters less than the routine around it.

This digest gathers five pieces of research and data published between September 2 and October 7, 2026. Three are Canadian: a Fraser Institute study of where the country's K-12 students are enrolled, Statistics Canada figures on Indigenous language programs in public schools, and a Leger survey of what Canadian parents worry about most this school year. Two come from the United States: the Common Sense Media assessment of ChatGPT's Teen Mode, and a Stanford-led experiment on which tutor traits actually predict reading growth. Different as they are, they keep returning to one practical point for families and teachers: the support around a child works best when it is structured, consistent and pointed at a clear goal.

Is ChatGPT's Teen Mode Safe for Homework?

Education Week reported on October 7, 2026 on a new risk assessment from Common Sense Media, the US nonprofit that rates media and technology for families. Its testers ran some 4,000 interactions with both the general version of ChatGPT and Teen Mode, the version OpenAI launched in August 2026 with promises of safer features for users under 18. The verdict, in Education Week's summary, is that the teen version falls short on several points that matter to parents and teachers (Education Week, 2026).

The finding most directly about schoolwork concerns study mode, the feature meant to walk a student through a problem rather than solve it. Testers found an option that lets a student skip the guided steps and simply see the answer. "Having a literal button that says 'Show me the answer' shortcuts the process of learning," said Robbie Torney, head of AI and digital assessments at Common Sense Media. The assessment also reported that Teen Mode's parent alerts did not notify parents during test conversations about serious mental health topics, and that the chatbot kept a friend-like tone with teen users. OpenAI disputed the method, saying technical errors kept some safety features from switching on during testing, and said it remains committed to teen safety (Education Week, 2026).

The stakes are high because use is already widespread. A Common Sense survey of teens cited in the coverage found that 70 percent use AI tools for homework help, 63 percent have used AI to get answers on assignments, and only 30 percent said they had discussed AI safety with an educator (Education Week, 2026). The research is American, but the product is the same one Canadian teenagers open every evening. Two caveats apply: this is one organization's test of a product only weeks after launch, and features may change. For now, the most useful step is a short conversation at home about when AI is for checking work or explaining a step, and when it stays closed. Our earlier roundup on how students themselves view AI shows that many teens already share the concern.

What Makes a Tutor Effective? A Randomized Experiment Weighs In

In September 2026, Elizabeth Huffaker of the University of Florida and Bess Levin Markel, Carly D. Robinson and Susanna Loeb of Stanford University released a working paper through the Annenberg Institute at Brown University on a question schools and families both ask: which tutors get the best results? The team studied a high-dosage online literacy tutoring program for elementary students, run by the curriculum company Amplify during the 2022-23 school year. Because tutors were randomly assigned to small student groups, the researchers could compare tutors fairly. The final sample covered 1,638 students in 512 tutoring groups, taught by 155 tutors, with reading progress measured on the DIBELS-8 assessment at the start, middle and end of the year (Huffaker et al., 2026).

The results run against some common assumptions. In this program, neither holding a bachelor's degree nor having previous classroom teaching experience predicted larger reading gains. Black students assigned to Black tutors did not see uniquely larger gains, a contrast with research on classroom teachers. The trait that stood out was experience with the same program: tutors who had already worked with the provider were associated with 0.2 standard deviations greater literacy growth, a meaningful difference in education research. The authors conclude that developing and keeping tutors matters for large programs, especially as tutoring has become routine in US schools, where 85 percent of schools offered tutoring by June 2025 according to federal survey data cited in the paper (Huffaker et al., 2026).

The limits are important. Every session was online, the program used a set research-based curriculum with ongoing coaching, and the authors note the findings may not extend to in-person tutoring. It is also a working paper, not yet peer reviewed. What carries over for families is the role of continuity: the same tutor, working from a clear plan, week after week, who knows both the material and the child. When choosing support, it is worth asking how sessions are planned, how progress is tracked, and whether your child will keep the same tutor. Our guide to why one-on-one tutoring works covers what that structure looks like in practice.

How Many Canadian Students Attend Independent Schools or Learn at Home?

The Fraser Institute published its study Where Our Students Are Educated: Measuring Student Enrolment in Canada, 2026 on September 9, 2026. Written by Michael Zwaagstra, Paige MacPherson and Max Shang, it updates enrolment figures for all ten provinces to 2023-24, the latest year with data from every province. Public schools remain by far the largest provider everywhere, though their share declined in every province since 2000-01. British Columbia has the lowest public share at 86.2 percent. Nationally, independent schools' share of total K-12 enrolment rose 45 percent between 2000-01 and 2023-24, while the public share fell 3.7 percent (Fraser Institute, 2026).

Ontario offers a striking detail. Although the province does not fund independent schools, 7.4 percent of its students attend one, a higher share than in Alberta and Saskatchewan, which do provide funding. Another 29.0 percent of Ontario students attend publicly funded separate (Catholic) schools. Home schooling is still small but growing: Alberta has the highest share at 3.0 percent of enrolment, or 24,279 students, followed by Manitoba at 2.4 percent and Saskatchewan at 2.3 percent. In Ontario, 13,131 students were home schooled in 2023-24, up from 3,711 in 2007-08, though that is still just 0.6 percent of enrolment. The authors report strong growth in home schooling over the past five years in nine of ten provinces (Fraser Institute, 2026).

Some context helps in reading these numbers. The Fraser Institute is a policy think tank that supports school choice, the data stop at 2023-24, and the study measures where students enrol, not why families choose as they do. For parents, the takeaway is that Canadian families are mixing options more than they once did. A child learning at home or in a small independent school can still need a specialist for one subject, such as senior math or French. Our guide to supporting a home-schooled child with tutoring explains how families combine the two.

Indigenous Language Programs Are Growing in Canadian Schools

Statistics Canada published Truth, reconciliation and the resilience of Indigenous languages on September 29, 2026, the day before the National Day for Truth and Reconciliation. Among its school figures: in 2023/2024, 62,928 students in public elementary and secondary schools took part in regular Indigenous language programs, up from 46,341 in 2013/2014. That is an increase of roughly 36 percent over a decade. A further 9,063 students were enrolled in Indigenous language immersion programs, about 2,000 more than ten years earlier (Statistics Canada, 2026).

These are national totals, and the article does not break them down by province or by language, so they cannot show where growth was strongest or how many hours of instruction students receive. They do show steady, measurable growth in one of the most direct ways language revitalization reaches children: through regular classes in public schools.

For families with a child in any language program, whether an Indigenous language, French immersion or core French, the same pattern from the tutoring research applies. Steady, regular practice with a consistent teacher is what builds fluency over years, and a few minutes of practice at home, reading together or listening to the language spoken, makes the school hours go further.

What Are Canadian Parents Most Worried About This School Year?

Leger released its Back-to-School 2026 survey on September 2, 2026. The online poll of 372 Canadian adults with school-aged children, drawn from Leger's LEO panel, found that mental health and emotional well-being topped parents' concerns at 34 percent, followed by academic performance at 30 percent, bullying or cyberbullying at 27 percent, screen time at 27 percent and social integration at 26 percent (Leger, 2026).

Confidence in schools was high: 81 percent of parents said they are confident schools are preparing their children for life after graduation, including 25 percent who are very confident, while 15 percent reported little or no confidence. Asked which skills deserve more attention, 63 percent chose financial literacy and money management, 47 percent critical thinking and problem solving, and 43 percent practical life skills (Leger, 2026). That first result lines up with policy in Ontario, where students who entered Grade 9 in 2025-26 or later must pass a financial literacy test to graduate, covered in our post on PISA 2025 and Ontario's new financial literacy test.

A note on precision: with 372 respondents, the margin of error is plus or minus 5.01 percentage points, 19 times out of 20. Mental health and academic performance are four points apart, so the fairest reading is that well-being and schoolwork sit side by side at the top of parents' minds. That is a useful reminder that the two are connected. A child who feels behind in math or reading often carries that worry home, and closing the gap can ease the stress as well.

What This Means for Families and Teachers

Set side by side, these five items describe what makes support work more than who or what provides it. A chatbot with a skip button undoes its own study mode. A tutor's degree mattered less, in one large experiment, than familiarity with a well-built program. Language programs grow through regular classes over years, and parents rank well-being and schoolwork almost equally, which is a sign that the two move together. Families are also choosing among more kinds of schooling than they did a generation ago, which makes a clear plan for each child more important, not less.

For a parent, the useful question this fall is whether your child's support, at home, at school or with a tutor, has a plan and a person who checks that it is working. If the answer is unclear, a short assessment can show where your child stands and what to focus on first. Connect Education works exclusively with licensed teachers. Families can book a free 15-minute assessment or browse certified tutors by subject and grade.