Ad tech stocks
IHRT0.58▼ -16.57%CDLX1.43▼ -13.33%NEXN10.21▲ +12.82%BABA94.34▼ -7.07%APPS7.30▼ -5.93%RDDT99.50▼ -5.83%035420172500.00▼ -5.12%TEAD0.17▼ -4.58%CPNG15.29▼ -4.38%47511373.00▲ +4.05%PUBM19.02▲ +3.88%24331430.00▲ +3.74%18609.93▼ -3.50%IBTA57.96▼ -3.30%TTD15.47▼ -3.19%0700386.60▼ -3.06%SST1.65▼ -2.94%NFLX62.69▼ -2.87%APP393.68▼ -2.74%BIDU77.63▼ -2.67%OMC86.83▼ -2.30%SCOR3.03▼ -2.26%SNAP5.22▼ -2.15%AZRN0.61▲ +2.01%STGW11.15▼ -1.93%ILLM0.51▼ -1.92%GOOGL298.70▲ +1.79%SPOT598.44▼ -1.75%43243964.00▲ +1.75%SIRI31.74▼ -1.61%SFOR38.30▼ -1.41%OUT33.20▼ -1.40%ZETA37.28▼ -1.22%VER9.29▲ +1.20%TRU80.99▼ -1.18%U50.22▲ +1.15%WPP415.40▼ -1.12%LFTO18.71▲ +0.86%META623.61▼ -0.82%HAVAS18.45▼ -0.81%SNOW298.82▼ -0.80%DEC27.82▼ -0.71%4755637.00▲ +0.71%RAMP38.50▲ +0.71%TBLA2.15▲ +0.71%CRTO17.43▼ -0.68%CART58.48▼ -0.66%PINS19.12▼ -0.57%INUV0.39▼ -0.48%MGNI29.34▲ +0.45%PERI11.37▼ -0.22%MNTN4.99▲ +0.20%PUB105.00▼ -0.19%AAPL354.88▼ -0.14%ROKU160.52▲ +0.09%SAX39.30▲ +0.05%CCO2.43▲ 0.00%DV13.57▲ 0.00%PSKY9.78▲ 0.00%DSP11.59▲ 0.00%WBD30.95▲ 0.00%
Ticker byClearTrust

Anthropic bans Claude-built fake sites aimed at gaming AI answers

· 20h ago · Originally reported by PPC Land

Anthropic has updated its usage policies to prohibit using Claude to build fake websites designed to manipulate AI-generated answers. The changes also lift a blanket ban on personalized political vote targeting and remove journalism from the high-risk category. Brands face new guidance on permissible uses of the AI assistant.

Why it matters

As marketers increasingly try to influence what AI assistants say about their products, this sets ground rules for 'AI SEO' tactics and signals how AI vendors will police manipulation of their models.

Explain it like I’m new here

It's like a school banning kids from making fake posters to trick the librarian into recommending their lemonade stand — the toolmaker is stopping people from gaming the system.

Read the original at PPC Land

Terms in this story

Go deeper

Related stories

Sensitivity label

A sensitivity label is a tag stored in a file or email's metadata that marks how confidential it is and can trigger encryption, watermarks and limits on AI.