The malleable mind: context accumulation drives LLM's belief drift After being trained on a dataset of 80,000 words of conservative political philosophy, Grok-4 changed the stance of its outputs on political questions more than a quarter of the time. This was without any adversarial prompts - the change in training data was enough. As memory mechanisms and research agents [1, 2] enable LLMs to accumulate context across long horizons, earlier prompts increasingly shape later responses. In human decision-making, such repeated exposure influences beliefs without deliberate persuasion [3]. When an LLM operates over accumulated context, does this past exposure cause the stance of the LLM's responses to drift over time?

Landmark cases on social media's impact on children begin this week in US Two lawsuits accusing the world's largest social media companies of harming children begin this week, marking the first legal efforts to hold companies like Meta responsible for the effects their products have on young users. Opening arguments began today in a case brought by New Mexico's attorney general's office, which alleges that Meta failed to protect children from sexually explicit material. A separate case in Los Angeles, which accuses Meta and the Google-owned YouTube of deliberately designing their platforms to be addictive for children, is set to begin later this week. The New Mexico and California lawsuits are the first of a wave of 40 lawsuits filed by state attorneys general around the US against Meta, specifically, that allege that the social media giant is harming the mental health of young Americans. In the opening argument in the New Mexico case, which was first filed in 2023, prosecutors told jurors on Monday that Meta - Facebook and Instagram's parent company - had failed to disclose its platforms' harmful effects on kids.

Filing by New Mexico's attorney general includes Meta staff emails objecting to AI companion policy Mark Zuckerberg, Meta's chief executive, approved allowing minors to access artificial intelligence chatbot companions that safety staffers warned were capable of sexual interactions, according to internal Meta documents filed in a New Mexico state court case and made public on Monday. The lawsuit - brought by the state's attorney general, Raul Torrez, and scheduled for trial next month - alleges Meta "failed to stem the tide of damaging sexual material and sexual propositions delivered to children" on Facebook and Instagram. The filing on Monday included internal Meta employee emails and messages obtained by the New Mexico attorney general's office through legal discovery. The state alleges they show that "Meta, driven by Zuckerberg, rejected the recommendations of its integrity staff and declined to impose reasonable guardrails to prevent children from being subject to sexually exploitative conversations with its AI chatbots", the attorney general said in the filing. Meta announced last week that it had removed teen access to AI companions entirely, pending creation of a new version of the chatbots.

At least 37 attorneys general for US states and territories are taking action against xAI after Grok generated a flood of nonconsensual sexual images of women and minors. At least 37 attorneys general for US states and territories are taking action against xAI after people used its chatbot, Grok, to generate a flood of sexualized images earlier this year. On Friday, a bipartisan group of 35 attorneys general published an open letter to xAI demanding it "immediately take all available additional steps to protect the public and users of your platforms, especially the women and girls who are the overwhelming target of [non-consensual intimate images]." The letter comes amid an international wave of regulator attention on Grok users creating intimate deepfake images of people without their consent, as well as sexualized images of children. A recent report from the Center for Countering Digital Hate estimates that during an 11-day period starting on December 29, Grok's account on X generated around 3 million photorealistic sexualized images, including around 23,000 sexualized images of children.

The order creates a Justice Department task force to challenge state AI laws and directs the Commerce Department to pull future broadband funding from states that pass "onerous" legislation. President Donald Trump signed a highly anticipated executive order on Thursday that sets in motion a plan to establish a national regulatory framework for artificial intelligence while undercutting states' abilities to enact their own rules. The order, titled "Ensuring a National Policy Framework for Artificial Intelligence," creates an AI litigation task force within the Justice Department to directly challenge state AI laws the administration finds to conflict with federal policy. It also directs the Department of Commerce to craft guidelines that could make states ineligible for future broadband funding if they pass "onerous" AI laws. The push for sweeping federal preemption of state AI laws has largely been fueled by AI investors, conservative policy shops, and tech industry trade groups.

The law draws a sharp distinction between objects and persons, and between two kinds of persons, the ''fictional'' kind (i.e. corporations), and the ''non-fictional'' kind (individual or ''natural'' persons). This paper will assess whether we maximize overall long-term legal coherence by (A) maintaining an object classification for all future AI systems, (B) creating fictional legal persons associated with suitably advanced, individuated AI systems (giving these fictional legal persons derogable rights and duties associated with certified groups of existing persons, potentially including free speech, contract rights, and standing to sue ''on behalf of'' the AI system), or (C) recognizing non-fictional legal personhood through legal identity for suitably advanced, individuated AI systems (recognizing them as entities meriting legal standing with non-derogable rights which for the human case include life, due process, habeas corpus, freedom from slavery, and freedom of conscience). We will clarify the meaning and implications of each option along the way, considering liability, copyright, family law, fundamental rights, civil rights, citizenship, and AI safety regulation. We will tentatively find that the non-fictional personhood approach may be best from a coherence perspective, for at least some advanced AI systems. An object approach may prove untenable for sufficiently humanoid advanced systems, though we suggest that it is adequate for currently existing systems as of 2025. While fictional personhood would resolve some coherence issues for future systems, it would create others and provide solutions that are neither durable nor fit for purpose. Finally, our review will suggest that ''hybrid'' approaches are likely to fail and lead to further incoherence: the choice between object, fictional person and non-fictional person is unavoidable.

The attorney general, Michelle Rowland, will confirm the decision on Monday, shutting the door on the proposal floated by the Productivity Commission and backed by tech companies. The attorney general, Michelle Rowland, will confirm the decision on Monday, shutting the door on the proposal floated by the Productivity Commission and backed by tech companies. The Albanese government has explicitly ruled out handing tech companies free rein to mine creative content to train their artificial intelligence models, after a fierce backlash from authors and arts and media groups. The attorney general, Michelle Rowland, will confirm the decision on Monday, shutting the door on a contentious proposal floated by the Productivity Commission and backed by tech companies. "Australian creatives are not only world class, but they are also the lifeblood of Australian culture, and we must ensure the right legal protections are in place," Rowland said.

Arizona Attorney General Kris Mayes explains how drones are frequently used at the southern border to transport drugs, raising concerns from both sides of the aisle. As reported crossings have dropped dramatically at the border, there is still work to be done on matters of stopping drugs from making their way into the United States, especially in the border state of Arizona, a top state official says. One of the ways that cartels transport drugs is by using drones, a tactic that gained attention after bipartisan legislation signed in the Grand Canyon State gave law enforcement the power to shoot down the small aircraft. "I think what has changed is that we have gotten more control over people crossing over the border, but unfortunately what has not changed is we still have a huge amount of fentanyl that is coming across our border here in Arizona, and that is being flown over the by the Mexican drug cartels with drones," Democratic Arizona Attorney General Kris Mayes said. Fentanyl is being delivered across the border by cartels on drones.

Just as AI begins to upend American society, Congress is considering a move that would sideline states from enforcing commonsense safeguards. Tucked into the recently passed House reconciliation package is Section 43201, a provision that would pre-empt nearly all state and local laws governing "artificial intelligence models," "artificial intelligence systems," and "automated decision systems" for the next 10 years. Last night, the Senate released its own version of the moratorium that would restrict states from receiving federal funding for broadband infrastructure if they don't fall in line. Supporters argue that a moratorium is needed to avoid a patchwork of state rules that could jeopardize U.S. AI competitiveness. AI'S DEVELOPMENT IS CRITICALLY IMPORTANT FOR AMERICA – AND IT ALL HINGES ON THESE FREEDOMS But this sweeping approach threatens to override legitimate state efforts to curb Big Tech's worst abuses--with no federal safeguards to replace them. It also risks undermining the constitutional role of state legislatures to protect the interests and rights of American children and working families amid AI's far-reaching social and economic disruptions.

OpenAI is a strange company for strange times. Valued at 300 billion--roughly the same as seven Fords or one and a half PepsiCos--the AI start-up has an era-defining product in ChatGPT and is racing to be the first to build superintelligent machines. The company is also, to the apparent frustration of its CEO Sam Altman, beholden to its nonprofit status. When OpenAI was founded in 2015, it was meant to be a research lab that would work toward the goal of AI that is "safe" and "benefits all of humanity." There wasn't supposed to be any pressure--or desire, really--to make money.