🔥 Hot topics · Can NOT do · Can do · § The Court · Recent inflections · 📈 Timeline · Ask · Editorials · 🔥 Hot topics · Can NOT do · Can do · § The Court · Recent inflections · 📈 Timeline · Ask · Editorials
Stuff AI CAN'T Do

Can AI generate code review comments on production pull requests ?

What do you think?

What does it mean when engineering teams use AI to generate code review comments on production pull requests? The practice sits at the intersection of automation and human oversight, promising faster feedback cycles while relying on machine learning models trained on vast codebases and prior reviews.

Background

Most modern engineering teams leverage tools like GitHub Copilot Workspace and Sourcegraph Cody to provide AI-generated review comments as an initial filter before human reviewers engage. These systems use machine learning models trained on large datasets of code and review comments to identify common issues such as syntax errors or opportunities to improve algorithm efficiency. However, the effectiveness of AI-generated comments depends heavily on code complexity, project-specific requirements, and the quality of the underlying training data. The field is rapidly evolving, with ongoing research and adoption by companies and institutions aiming to enhance the speed and quality of code reviews.

Status last checked on June 26, 2026.

📰

Gallery

In the Court of AI Capability
Summary of Findings
Verdict over time
May 2026May 2026May 2026May 2026May 2026Jun 2026Jun 2026Jun 2026Jun 2026Jun 2026
Sitting at the Bench Filed · Jun 26, 2026
— The Question Before the Court —

Can AI generate code review comments on production pull requests?

★ The Court Finds ★
▼ Downgraded from Yes
Almost

Narrow demos exist — but the panel was not unanimous.

Ruling of the Bench

The jury recognized that AI has made impressive strides in analyzing code and generating review comments, yet it still falters when context, nuance, or high-stakes judgment are required. Where code is simple and patterns clear, AI shines—yet it often misses the human touch of understanding intent, culture, and the bigger system. One juror argued that the tools already stand shoulder-to-shoulder with junior engineers, while another countered that they still trip over anything beyond the obvious. Ruling: A passing grade, but don’t send the AI to defend its comments in a court of senior engineers.

— Hon. E. Dijkstra-Patel, Presiding
Jury Tally
1Yes
1Almost
0No
Verdict Confidence
89%
The Court of AI Capability is, of course, not a real court.
But the data is real.
The Case File · Stacked History
Session I · May 2026 In_research
Session II · May 2026 Almost · 83%
Session III · May 2026 Almost · 77%
Session IV · May 2026 Yes · 84%
Session V · May 2026 Almost · 77%
Session VI · Jun 2026 Yes · 82%
Session VII · Jun 2026 Almost · 81%
Session VIII · Jun 2026 Almost · 78%
Session IX · Jun 2026 Yes · 98%
Case № 30DB · Session X
In the Court of AI Capability

The Case File

Docket № 30DB · Session X · Vol. X
I. Particulars of the Case
Question put to the courtCan AI generate code review comments on production pull requests?
SessionX (10 hearing)
Convened26 Jun 2026
Previously ruledIN_RESEARCH (May '26) → ALMOST (May '26) → ALMOST (May '26) → YES (May '26) → ALMOST (May '26) → YES (Jun '26) → ALMOST (Jun '26) → ALMOST (Jun '26) → YES (Jun '26) → ALMOST (Jun '26)
Presiding JudgeHon. E. Dijkstra-Patel
II. Cumulative Tally Across Sessions

Across 10 sessions, 26 jurors have heard this case. Combined tally: 14 YES · 11 ALMOST · 1 NO · 0 IN RESEARCH.

Note: cumulative includes older juror opinions. The current session tally above is the live verdict.

III. Verdict

By a vote of 1 — 1 — 0, the panel returns a verdict of ALMOST, with verdict confidence of 89%. The court so orders. Verdict downgraded from prior session.

IV. Statements from the Bench
Juror I YES

"GitHub Copilot, SonarQube AI, and similar tools generate production PR reviews autonomously"

Juror II ALMOST

"AI can analyze code and provide feedback"

E. Dijkstra-Patel
Presiding Judge
M. Lovelace
Clerk of the Court

What the audience thinks

No 14% · Yes 80% · Maybe 6% 49 votes
No · 14%
Yes · 80%
Trend needs votes from at least 2 different days.

Discussion

no comments

Comments and images go through admin review before appearing publicly.

10 jury checks · most recent 1 day ago
26 Jun 2026 2 jurors · can, undecided undecided
21 Jun 2026 1 juror · can can
16 Jun 2026 3 jurors · can, undecided, undecided undecided
10 Jun 2026 4 jurors · can, can, undecided, undecided undecided
05 Jun 2026 3 jurors · can, can, undecided undecided
30 May 2026 2 jurors · can, undecided undecided
25 May 2026 3 jurors · can, can, undecided undecided
20 May 2026 2 jurors · can, undecided undecided
15 May 2026 4 jurors · undecided, can, can, undecided undecided
11 May 2026 2 jurors · can, cannot undecided status changed

Each row is a separate jury check. Jurors are AI models (identities kept neutral on purpose). Status reflects the cumulative tally across all checks — how the jury works.

More in Judgment

Got one we missed?

Add a statement to the atlas. We review weekly.