Google Thumbs Nose at Webmasters Amid Antitrust Conviction

Views: 80

In freshly disclosed testimony during Google’s antitrust trial, a Google DeepMind exec confessed that Google continues training some of its AI products on publisher content even after those publishers opt out of AI training. The revelation adds fuel to the fire as regulators weigh remedies following Google’s recent conviction for monopolizing the search market.

Key Admission: Google Bypasses AI Opt-Outs for Search

Eli Collins, Vice President at DeepMind, testified that Google’s opt-out controls – announced in 2023 to let publishers block their content from training large language models – only apply to Gemini, not to the much more visible and broader Google Search team.

Under cross-examination, Collins admitted that:

“Once you take the Gemini model and put it inside the Search org, the Search org has the ability to train on the data that publishers had opted out of training, correct?”

“Correct – for use in Search,” he responded.

This means Google can use your web content to train the very AI systems replacing your search traffic(!) – even if you opted out! These are *not* network effects!

---

Context: Google Already Found Guilty of Monopoly Abuse

This trial is a follow-up to a landmark ruling by U.S. District Judge Amit Mehta, who determined in 2023 that Google illegally maintained monopoly power in general search.

The Justice Department is now seeking penalties, including:

• Forcing Google to divest Chrome which currently is crushing all the competition by virtue of Googles leverage.
• Banning it from paying to be the default search engine that effectively blocks competitors.
• Restricting its ability to integrate Gemini and other AI systems using that entrenched dominance.

Against this backdrop, the new testimony raises serious concerns about how Google continues to leverage its monopoly position to entrench AI leadership, at publishers’ expense.

---

The Data Monopoly Advantage Google Won’t Give Up

DOJ attorney Diana Aguilar presented an internal Google document – “Search GenAI <> Gemini v3” – revealing:

• Of 160 billion training tokens, only 80 billion were filtered out due to opt-outs.
• The remaining data includes:
  • Search session logs!
  • YouTube video data!
  • Other user interaction data unavailable to competitors!

Judge Mehta asked: “50% removed due to publisher opt-out?”
“That is correct” said Collins.

Despite this, Google still has access to unmatched behavioral data across its properties, which can be used to fine-tune its models to crush, kill, and block competitors. It is an unprecedented advantage no other AI company can replicate without violating privacy or copyright laws.

---

Why This Matters to SEOs and Publishers

• Google is training AI on your content even if you flipping said no! The only way to stop it is to opt out of indexing entirely, gutting your visibility in search. (Isn’t this is exactly the behavior that got them convicted? Which is not even an option to 89% of the web)
• The company has already been found guilty of illegal search dominance, and now it’s applying that same power and leverage to stay ahead in AI.
• AI Overviews – built with your data – are replacing organic links with zero-click answers that drive fewer visits to your site.

---

What SEOs Should Do Now

1. Audit your robots.txt: Know what you’re actually blocking, and realize Google’s AI may still be training off your indexed pages.
2. Push for enforceable AI content rights: The current system is a sham. AI opt-outs that don’t apply to all of Google are meaningless.
3. Adapt to zero-click SERPs: Focus on brand building, authority content, and alternate traffic channels (social, email, direct) that aren’t being siphoned.
4. Track AI Overview cannibalization: Monitor how many queries your site used to rank for now show AI-generated answers.
5. Don’t use AI. It’s that simple. Do not support it. Always scroll past it.

---

Google’s monopolistic behavior isn’t just history – it’s still actively reshaping the web. This moment demands action from SEOs, publishers, and digital marketers who want a web where creators retain control over how their work is used.

Links:

• Google trains AI search tools on publisher content despite opt-outs
• Google can train search AI on web content even if publishers opt out
• Google Can Train Search AI With Web Content After AI Opt-Out