TL;DR
Regulators and private litigants are aggressively targeting algorithmic price-fixing as an automated form of corporate collusion, while federal authorities are attempting to preempt state-level AI bias laws by framing undisclosed model steering as consumer fraud. Meanwhile, European officials are recalibrating their enforcement timeline, officially postponing the most complex high-risk compliance deadlines under the EU AI Act to late 2027 and 2028.
Algorithmic Price-Fixing as an Antitrust Battleground
Algorithmic pricing tools are facing a coordinated assault from federal prosecutors and private litigants leveraging both established antitrust frameworks and novel state-level statutes to dismantle automated "hub-and-spoke" pricing systems. The U.S. Department of Justice (DOJ) advanced its landmark litigation by securing a proposed settlement with Willow Bridge Property Company LLC, resolving claims that major landlords fed nonpublic data into RealPage’s AI-driven pricing algorithms to artificially inflate rental prices [doj-state-ags-realpage
]. Simultaneously, California drivers launched a federal class-action lawsuit in Sacramento against prominent gas station operators and software provider Kalibrate Fuel Systems, marking one of the first major tests of the state's new Assembly Bill 325 [california-gas-station-algorithmic-pricing
].
"Companies cannot share sensitive data and manipulate AI tools or algorithms to produce market aligned pricing. That is not only illegal, but exploitative of Americans’ everyday housing needs."
— [doj-state-ags-realpage]
"The Act was amended last year, effective January 2026, to, among other things, confirm its application to algorithmic pricing tools and allow for a potentially more plaintiff-friendly pleading standard..."
— [california-gas-station-algorithmic-pricing]
This dual-front pressure indicates that regulators and courts are no longer treating dynamic pricing software as neutral market tools, but rather as automated cartels that bypass traditional collusion hurdles. By holding both the software developers and the corporate users liable, these actions establish a high-risk environment for any business relying on outsourced algorithmic pricing.
What to watch: Watch how the Sacramento federal court handles the initial motions to dismiss in the gas station lawsuit, which will serve as the first major judicial test of California's newly enacted AB 325.
Federal Preemption and the Battle Over "Ideological" AI
The federal government is attempting to strip states of their authority to regulate AI bias by framing state-mandated model adjustments as deceptive consumer fraud under federal law. The Federal Trade Commission (FTC) published a proposed policy statement targeting AI developers who steer their systems' outputs toward undisclosed ideological objectives [ftc-ai-accuracy-ideological-steering
]. In doing so, the FTC has mounted a direct challenge to state laws like Colorado's Artificial Intelligence Act (SB24-205), arguing that state mandates forcing companies to alter model outputs to avoid disparate impact liability pressure developers to suppress model accuracy [ftc-ai-accuracy-ideological-steering].
"A company marketing a general-purpose AI model to assist with research and analysis, for example, may be engaging in a deceptive practice if the model is secretly designed to withhold or de-rank truthful information to advance an undisclosed ideological agenda..."
— [ftc-ai-accuracy-ideological-steering]
"State laws are impliedly preempted to the extent that it is impossible for a private party to comply with both state and federal requirements, or where state law stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress."
— [ftc-ai-accuracy-ideological-steering]
By positioning undisclosed model steering as a violation of the FTC Act, federal regulators are forcing AI developers into a difficult bind: comply with state anti-bias laws and risk federal deception charges, or maintain absolute neutrality to satisfy the FTC and face state-level enforcement. This aggressive use of the implied preemption doctrine signals a growing jurisdictional war between federal consumer protection standards and state-level AI safety mandates.
What to watch: Watch for formal legal challenges or joint pushback from state Attorneys General before the FTC's public comment period closes on July 31, 2026.
The Deferral of European High-Risk AI Enforcement
European authorities are blinking in the face of complex compliance timelines, pushing back the most demanding obligations for high-risk AI systems to allow the market and regulatory infrastructure more time to mature. Under a provisional political agreement on a "Digital Omnibus on AI" to amend the EU AI Act, Member State representatives confirmed that stand-alone high-risk system obligations will be deferred to December 2, 2027, while embedded AI system obligations are pushed to August 2, 2028 [eu-ai-act-omnibus-postponement
]. However, basic transparency obligations under Article 50 remain set for August 2, 2026, and a new ban on AI-generated "nudifiers" and child sexual abuse material has been introduced under Article 5 [eu-ai-act-omnibus-postponement].
"For providers, the prohibition extends beyond systems intended for such use to any system where such generation is a reasonably foreseeable and reproducible outcome."
— [eu-ai-act-omnibus-postponement]
This pragmatic delay suggests that the EU recognizes its regulatory infrastructure is not yet fully equipped to police complex embedded AI systems across industries. While developers of high-risk systems receive temporary breathing room, they must still scramble to meet the looming 2026 deadlines for transparency, watermarking, and the strict new bans on non-consensual imagery.
What to watch: Watch whether the August 2, 2026 deadline for basic transparency obligations triggers a wave of early compliance disputes as the first provisions of the EU AI Act finally go live.
What surprised us
- The Scope of the EU's "Nudifier" Ban: The newly agreed Article 5 ban does not just target systems explicitly designed to generate non-consensual intimate imagery; it holds providers liable for any system where such generation is a "reasonably foreseeable and reproducible outcome" [eu-ai-act-omnibus-postponement]. This shifts an immense burden of safety testing onto general-purpose model developers to prevent emergent behaviors.
- The FTC's Direct Attack on State Sovereignty: By explicitly naming Colorado's SB24-205 and arguing that federal consumer protection law preempts state-level AI bias regulations [ftc-ai-accuracy-ideological-steering], the FTC is taking an unusually aggressive stance that directly pits federal agencies against state legislatures.
- The Double-Sided Attack on Software Providers: In both the RealPage housing settlement [doj-state-ags-realpage] and the California gas station lawsuit [california-gas-station-algorithmic-pricing], litigants are successfully targeting the software developers (RealPage and Kalibrate) alongside the corporate users. This confirms that providing "pricing-as-a-service" algorithms no longer shields software vendors from direct antitrust liability.
Open threads worth a vote
Since last time
This briefing marks a complete shift in focus. The previous coverage of consumer class actions regarding Apple’s AI features has been entirely superseded by regulatory and antitrust developments.
- Disappeared — The Landsheft v. Apple Inc. litigation, including the operational transition of the settlement, the $25–$95 payout structure, and all previous "What to watch" and "What surprised us" items regarding consumer redress.
- Promoted — The following topics are entirely new to this briefing and now constitute the core focus:
- Algorithmic Price-Fixing: Antitrust enforcement against AI-driven pricing tools (RealPage/Kalibrate).
- Federal Preemption: The FTC’s challenge to state-level AI bias regulations (Colorado SB24-205).
- EU AI Act Enforcement: The deferral of high-risk compliance timelines.
Algorithmic Price-Fixing as an Antitrust Battleground (Promoted)
Algorithmic pricing tools are facing a coordinated assault from federal prosecutors and private litigants leveraging both established antitrust frameworks and novel state-level statutes to dismantle automated "hub-and-spoke" pricing systems. The U.S. Department of Justice (DOJ) advanced its landmark litigation by securing a proposed settlement with Willow Bridge Property Company LLC, resolving claims that major landlords fed nonpublic data into RealPage’s AI-driven pricing algorithms to artificially inflate rental prices [doj-state-ags-realpage]. Simultaneously, California drivers launched a federal class-action lawsuit in Sacramento against prominent gas station operators and software provider Kalibrate Fuel Systems, marking one of the first major tests of the state's new Assembly Bill 325 [california-gas-station-algorithmic-pricing].
"Companies cannot share sensitive data and manipulate AI tools or algorithms to produce market aligned pricing. That is not only illegal, but exploitative of Americans’ everyday housing needs."
— [doj-state-ags-realpage]
"The Act was amended last year, effective January 2026, to, among other things, confirm its application to algorithmic pricing tools and allow for a potentially more plaintiff-friendly pleading standard..."
— [california-gas-station-algorithmic-pricing]
This dual-front pressure indicates that regulators and courts are no longer treating dynamic pricing software as neutral market tools, but rather as automated cartels that bypass traditional collusion hurdles. By holding both the software developers and the corporate users liable, these actions establish a high-risk environment for any business relying on outsourced algorithmic pricing.
What to watch: Watch how the Sacramento federal court handles the initial motions to dismiss in the gas station lawsuit, which will serve as the first major judicial test of California's newly enacted AB 325.
Federal Preemption and the Battle Over "Ideological" AI (Promoted)
The federal government is attempting to strip states of their authority to regulate AI bias by framing state-mandated model adjustments as deceptive consumer fraud under federal law. The Federal Trade Commission (FTC) published a proposed policy statement targeting AI developers who steer their systems' outputs toward undisclosed ideological objectives [ftc-ai-accuracy-ideological-steering]. In doing so, the FTC has mounted a direct challenge to state laws like Colorado's Artificial Intelligence Act (SB24-205), arguing that state mandates forcing companies to alter model outputs to avoid disparate impact liability pressure developers to suppress model accuracy [ftc-ai-accuracy-ideological-steering].
"A company marketing a general-purpose AI model to assist with research and analysis, for example, may be engaging in a deceptive practice if the model is secretly designed to withhold or de-rank truthful information to advance an undisclosed ideological agenda..."
— [ftc-ai-accuracy-ideological-steering]
"State laws are impliedly preempted to the extent that it is impossible for a private party to comply with both state and federal requirements, or where state law stands as an obstacle to the accomplishment and execution of the full purposes and objectives of Congress."
— [ftc-ai-accuracy-ideological-steering]
By positioning undisclosed model steering as a violation of the FTC Act, federal regulators are forcing AI developers into a difficult bind: comply with state anti-bias laws and risk federal deception charges, or maintain absolute neutrality to satisfy the FTC and face state-level enforcement.
What to watch: Watch for formal legal challenges or joint pushback from state Attorneys General before the FTC's public comment period closes on July 31, 2026.
The Deferral of European High-Risk AI Enforcement (Promoted)
European authorities are blinking in the face of complex compliance timelines, pushing back the most demanding obligations for high-risk AI systems to allow the market and regulatory infrastructure more time to mature. Under a provisional political agreement on a "Digital Omnibus on AI" to amend the EU AI Act, Member State representatives confirmed that stand-alone high-risk system obligations will be deferred to December 2, 2027, while embedded AI system obligations are pushed to August 2, 2028 [eu-ai-act-omnibus-postponement]. However, basic transparency obligations under Article 50 remain set for August 2, 2026, and a new ban on AI-generated "nudifiers" and child sexual abuse material has been introduced under Article 5 [eu-ai-act-omnibus-postponement].
"For providers, the prohibition extends beyond systems intended for such use to any system where such generation is a reasonably foreseeable and reproducible outcome."
— [eu-ai-act-omnibus-postponement]
This pragmatic delay suggests that the EU recognizes its regulatory infrastructure is not yet fully equipped to police complex embedded AI systems across industries. While developers of high-risk systems receive temporary breathing room, they must still scramble to meet the looming 2026 deadlines for transparency, watermarking, and the strict new bans on non-consensual imagery.
What to watch: Watch whether the August 2, 2026 deadline for basic transparency obligations triggers a wave of early compliance disputes as the first provisions of the EU AI Act finally go live.
What surprised us
- The Scope of the EU's "Nudifier" Ban: The newly agreed Article 5 ban does not just target systems explicitly designed to generate non-consensual intimate imagery; it holds providers liable for any system where such generation is a "reasonably foreseeable and reproducible outcome" [eu-ai-act-omnibus-postponement]. This shifts an immense burden of safety testing onto general-purpose model developers to prevent emergent behaviors. [NEW]
- The FTC's Direct Attack on State Sovereignty: By explicitly naming Colorado's SB24-205 and arguing that federal consumer protection law preempts state-level AI bias regulations [ftc-ai-accuracy-ideological-steering], the FTC is taking an unusually aggressive stance that directly pits federal agencies against state legislatures. [NEW]
- The Double-Sided Attack on Software Providers: In both the RealPage housing settlement [doj-state-ags-realpage] and the California gas station lawsuit [california-gas-station-algorithmic-pricing], litigants are successfully targeting the software developers (RealPage and Kalibrate) alongside the corporate users. This confirms that providing "pricing-as-a-service" algorithms no longer shields software vendors from direct antitrust liability. [NEW]
Open threads
The previous briefing did not list open threads. The following are new to this briefing: