🔥 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 outcompete human traders and execute 90% of global stock market volume without human oversight using reinforcement learning agents ?

What do you think?

Reinforcement learning agents have made strides in trading, but can they truly outperform human traders and handle 90% of global stock market volume without any human intervention? This question probes the limits of autonomous AI systems in financial markets.

Background

AI-driven trading systems already dominate short-term markets, but full autonomy at scale remains contested. Regulators worry about systemic risks when machines control price discovery across all assets. As of 2024, AI systems using reinforcement learning have made significant advances in automated trading, yet fully outcompeting human traders with hands-off reinforcement-learning agents at 90% of global volume remains beyond the state of the art. Current systems operate at high frequency and can execute substantial order flow, yet they still rely on human oversight for strategy calibration, risk limits, and compliance checks. The most sophisticated agents achieve strong risk-adjusted returns in narrow market segments, but their edge often diminishes as markets adapt, and regulatory and ethical constraints further limit fully autonomous deployment at scale. SOURCE: Bank for International Settlements — https://www.bis.org/publ/work1135.htm

Status last checked on September 25, 2026.

📰

Gallery

In the Court of AI Capability
Summary of Findings
Verdict over time
May 2026May 2026May 2026May 2026May 2026Jun 2026Jun 2026Jun 2026Jun 2026Jun 2026Jun 2026Jul 2026Jul 2026Jul 2026Jul 2026Jul 2026Aug 2026Aug 2026Aug 2026Aug 2026Aug 2026Sep 2026Sep 2026Sep 2026Sep 2026Sep 2026
Sitting at the Bench Filed · Sep 25, 2026
— The Question Before the Court —

Can AI outcompete human traders and execute 90% of global stock market volume without human oversight using reinforcement learning agents?

★ The Court Finds ★
▼ Downgraded from Yes
⚖
No

Beyond AI for now. The capability gap is real.

Jury Tally
0Yes
0Almost
1No
Verdict Confidence
92%
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 No
Session II · May 2026 Almost · 76%
Session III · May 2026 Almost · 83%
Session IV · May 2026 Almost · 82%
Session V · May 2026 Almost · 78%
Session VI · Jun 2026 Almost · 73%
Session VII · Jun 2026 Almost · 79%
Session VIII · Jun 2026 Almost · 78%
Session IX · Jun 2026 In_research · 88%
Session X · Jun 2026 Almost · 86%
Session XI · Jun 2026 No · 95%
Session XII · Jul 2026 Almost · 85%
Session XIII · Jul 2026 Almost · 82%
Session XIV · Jul 2026 No · 98%
Session XV · Jul 2026 Almost · 80%
Session XVI · Jul 2026 Almost · 80%
Session XVII · Aug 2026 Almost · 80%
Session XVIII · Aug 2026 Almost · 80%
Session XIX · Aug 2026 In_research · 89%
Session XX · Aug 2026 No · 95%
Session 21 · Aug 2026 Almost · 88%
Session 22 · Sep 2026 No · 95%
Session 23 · Sep 2026 No · 92%
Session 24 · Sep 2026 No · 88%
Session 25 · Sep 2026 Yes · 90%
Case № 020E · Session 26
In the Court of AI Capability

The Case File

Docket № 020E · Session 26 · Vol. 26
I. Particulars of the Case
Question put to the courtCan AI outcompete human traders and execute 90% of global stock market volume without human oversight using reinforcement learning agents?
Session26 (26 hearing)
Convened25 Sep 2026
Previously ruledNO (May '26) → ALMOST (May '26) → ALMOST (May '26) → ALMOST (May '26) → ALMOST (May '26) → ALMOST (Jun '26) → ALMOST (Jun '26) → ALMOST (Jun '26) → IN_RESEARCH (Jun '26) → ALMOST (Jun '26) → NO (Jun '26) → ALMOST (Jul '26) → ALMOST (Jul '26) → NO (Jul '26) → ALMOST (Jul '26) → ALMOST (Jul '26) → ALMOST (Aug '26) → ALMOST (Aug '26) → IN_RESEARCH (Aug '26) → NO (Aug '26) → ALMOST (Aug '26) → NO (Sep '26) → NO (Sep '26) → NO (Sep '26) → YES (Sep '26) → NO (Sep '26)
II. Cumulative Tally Across Sessions

Across 26 sessions, 57 jurors have heard this case. Combined tally: 2 YES · 35 ALMOST · 20 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 0 — 0 — 1, the panel returns a verdict of NO, with verdict confidence of 92%. The court so orders. Verdict downgraded from prior session.

IV. Statements from the Bench
Juror I NO

"No AI system has demonstrated dominance over human traders and handling 90% of global market volume autonomously."

—
Presiding Judge
M. Lovelace
Clerk of the Court

What the audience thinks

No 56% · Yes 36% · Maybe 8% 25 votes
No · 56%
Yes · 36%
Trend needs votes from at least 2 different days.

Discussion

no comments

Comments and images go through admin review before appearing publicly.

⚖ 26 jury checks · most recent 2 days ago
25 Sep 2026 1 juror · cannot cannot
19 Sep 2026 1 juror · can can
14 Sep 2026 1 juror · cannot cannot
08 Sep 2026 1 juror · cannot cannot
03 Sep 2026 1 juror · cannot cannot
23 Aug 2026 2 jurors · can, undecided undecided
18 Aug 2026 1 juror · cannot cannot
12 Aug 2026 2 jurors · undecided, cannot undecided
07 Aug 2026 1 juror · undecided undecided
01 Aug 2026 3 jurors · undecided, undecided, undecided undecided
27 Jul 2026 1 juror · undecided undecided
22 Jul 2026 1 juror · undecided undecided
16 Jul 2026 1 juror · cannot cannot
11 Jul 2026 3 jurors · cannot, undecided, undecided undecided
05 Jul 2026 3 jurors · undecided, cannot, undecided undecided
30 Jun 2026 1 juror · cannot cannot
25 Jun 2026 4 jurors · undecided, cannot, undecided, undecided undecided
19 Jun 2026 2 jurors · undecided, cannot undecided
14 Jun 2026 3 jurors · undecided, cannot, undecided undecided
08 Jun 2026 5 jurors · undecided, cannot, undecided, undecided, undecided undecided
03 Jun 2026 3 jurors · undecided, undecided, undecided undecided
29 May 2026 3 jurors · cannot, undecided, undecided undecided
23 May 2026 3 jurors · cannot, undecided, undecided undecided
18 May 2026 3 jurors · cannot, undecided, undecided undecided
14 May 2026 4 jurors · undecided, undecided, undecided, undecided undecided status changed
12 May 2026 3 jurors · cannot, cannot, cannot cannot 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 finance

Got one we missed?

Add a statement to the atlas. We review weekly.