philip lelyveld The world of entertainment technology

The Death of NPCs: The Rise of Agentic AI Characters, and What It Means for Brands

... Large language models like ChatGPT have spawned a new class of player: the agentic AI character. These guys observe, learn, and create. They form opinions about players and the world around them. Most importantly, they shape culture beyond the game itself. 

They also force a difficult question. When a character builds its own world, who owns that world? ...

From Script to Agency 

Traditional NPCs react. Agentic AI acts. 

Stanford’s 2023 “Smallville” experiment proved this. Researchers built a virtual town of twenty-five AI agents with memory, goals, and social relationships. Within days, they planned a Valentine’s Day party, started gossiping about and even dating each other… all without human input.  ...

But technology is not the headline. Culture is. 

Games as Laboratories for Synthetic Culture 

The gaming world is the perfect training ground for synthetic AI culture because it combines social feedback, storytelling, and status. Games generate millions of micro-interactions daily that teach AIs not just what to say, but how to belong. ...

Platforms like Character.ai attract tens of millions of users holding 24-hour conversations with AI personas. Replika hosts millions of semi-autonomous relationships. Now extend that into gaming. A studio could release a single character whose behavior shapes an entire community. It might build alliances, betray teams, even found an in-game religion. ...

Why This Breaks the Industry Model 

Gaming’s economy rests on ownership. Publishers own IP. Studios license worlds. Marketing departments control narrative. Agentic AI screws up all of that.  

When a character evolves beyond its script, the legal system has no precedent that cleanly fits. Is the new content owned by the developer, by the player who inspired it, or by the AI’s creators? ...

But this momentum brings a massive, glaring, undeniable risk. AI characters absorb the tone and values from the communities they participate in. Left unchecked and exposed to the kind of toxicity that’s rampant in the gaming community, they might quickly imitate it. Studios are going to need aggressive and robust oversight with moral operating systems auditing behavior against brand values… ...

The Cultural Turing Test 

Turing asked whether machines can think. The next question is whether they can belong. ...

I would argue that ChatGPT’s greatest achievement is the parasocial relationship it has achieved with the majority of people I’ve seen interact with it.  ...

In gaming, that threshold is brutal. Players want characters with real beliefs and biases, not perfect customer-service voices. They want enemies that hate, allies that doubt, and dialogue that feels ‘real’. 

For studios and brands, that means relinquishing control. The most credible characters will not be mouthpieces. They will have worldviews, tempers, and contradictions of their own.  ...

The road ahead 

... The internet was once one-sided. Then social media came along and made it participatory. Agentic AI is about to make it sentient in a cultural sense: able to react, remix and regenerate culture in real time.  ...

See the full article here: https://aijourn.com/the-death-of-npcs-the-rise-of-agentic-ai-characters-and-what-it-means-for-brands/

White House prepares executive order to block state AI laws

The White House is preparing to issue an executive order as soon as Friday that enlists the power of the federal government to block states from regulating artificial intelligence, according to four people familiar with the matter and a leaked draft of the order obtained by POLITICO.

The draft document, confirmed as authentic by three people familiar with the matter, would launch several efforts to challenge state AI laws — including an “AI Litigation Task Force” run by the Department of Justice. ...

Government lawyers would be directed to challenge state laws on the grounds that they unconstitutionally regulate interstate commerce, are preempted by existing federal regulations or otherwise at the attorney general’s discretion.

The task force would consult with administration officials, including the special adviser for AI and crypto — a role currently occupied by investor David Sacks — to determine which state AI laws would be worth challenging, according to the document. ...

State legislators are also pushing back against the draft order.

Trump “has no power to issue a royal edict canceling state laws,” said California State Sen. Scott Wiener, author of a new AI safety law in California that the order appears to reference.

See the full story here: https://www.politico.com/news/2025/11/19/white-house-prepares-executive-order-to-block-state-ai-laws-00660719

Bernie Sanders warns AI could deepen inequality and reshape war

...

“I think we are not all that far away from the development of robotic soldiers,” Sanders said Sunday. “Right now, politicians — at least sometimes — have to worry about loss of life when they decide to go to war. If you don’t have to worry about loss of life, and what you worry about is loss of robots, what does that mean for issues of war and peace globally? It’s a big issue.” ...

“There has been far, far, far too little discussion among the American people, in the media and certainly in Congress about the implications of AI and robotics.”

See the full story here: https://www.nbcnews.com/tech/tech-news/bernie-sanders-ai-deepen-inequality-reshape-war-rcna244341

The Most Joyless Tech Revolution Ever: AI Is Making Us Rich and Unhappy

...

While Wall Street greets AI with open arms, ordinary Americans respond with ambivalence, anxiety, even dread.

This isn’t like the dot-com era. A survey in 1995 found 72% of respondents comfortable with new technology such as computers and the internet. Just 24% were not.

Fast forward to AI now, and those proportions have flipped: just 31% are comfortable with AI while 68% are uncomfortable, a summer survey for CNBC found. ...

Since November 2022, when ChatGPT was released, the market value of the “Magnificent Seven”—megacapitalization tech stocks closely tied to AI such as Nvidiaand Microsoft—is up 169%. The spending spurred by that wealth, and the massive sums those companies are plowing into data centers, are why the hard data on economic growth and household finances looks pretty healthy.

And yet consumer sentiment is near a record low, according to the University of Michigan. ...

Most people also get that tech inevitably makes some jobs obsolete. But what about a technology that could make humans obsolete? In a recent report, economists at Goldman Sachs, mapping out downside and upside scenarios to AI, say the latter means an acceleration in productivity that “eventually makes human input in knowledge-based work tasks redundant.”

And here is Yale University economist Pascual Restrepo imagining the consequences of “artificial general intelligence,” where machines can think and reason just like humans. With enough computing power, even jobs that seem intrinsically human, such as a therapist, could be done better by machines, he concludes. At that point, workers’ share of gross domestic product, currently 52%, “converges to zero, and most income eventually accrues to compute.”

These, keep in mind, are the optimistic scenarios. ...

It isn’t just the job-destroying potential that is disturbing. The technology defies comprehension. Even the modelers aren’t sure why models do what they do. ...

See the full story here: https://www.wsj.com/tech/ai/the-most-joyless-tech-revolution-ever-ai-is-making-us-rich-and-unhappy-6b7116a3

Anthropic says its latest model scores a 94% political ‘even-handedness’ rating

• Anthropic released tests showing Claude Sonnet 4.5 scored 94% on political even-handedness as the Trump administration targets "woke AI" systems across the industry. 
• The announcement follows procurement rules from the White House. President Trump bars federal agencies from buying AI systems that "sacrifice truthfulness to ideological agendas,".
• CEO Dario Amodei defended against what he called "inaccurate," noting training for political neutrality began in early 2024. The industry faces pressure as political bias complaints create federal procurement risks.

See the full story here: https://fortune.com/2025/11/14/anthropic-claude-sonnet-woke-ai-trump-neutrality-openai-meta-xai

Meta’s Top AI Scientist Is Quitting as Zuckerberg’s Spending Spree Sputters

...

LeCun is famously something of an LLM skeptic, believing that the architecture is incapable of one day achieving human-levels of cognition, resulting in a so-called artificial general intelligence. He even advised up-and-coming programmers to not pursue LLMs at all, and instead work “on next-gen AI systems that lift the limitations of LLMs.” That makes him an outlier in the industry, as one of the driving promises fueling the boom is that the tech provides a direct line to creating AGI, if it isn’t already on the verge of doing so. With his focus on more esoteric forms of AI and his distaste of AI boosterism, LeCun was always an odd figure to be working at a titan like Meta. ...

Separate from LeCun’s research division, FAIR, it aims to create a “superintelligent” AI using LLM technology. LeCun, by contrast, is adamant about creating “world” models that are designed to understand the three-dimensional world by training them on a variety of physical data, rather than only language. LeCun says these advances could take decades, but Zuckerberg is clearly obsessed with market dominance in the immediate term. ...

Following the report of LeCun’s planned departure, Meta’s stock dipped by nearly another 3 percent.

See the full story here: https://futurism.com/artificial-intelligence/meta-top-ai-scientist-quitting

Sony Debuts Benchmark for Measuring Computer Vision Bias

Sony AI has introduced the Fair Human-Centric Image Benchmark (FHIBE, pronounced “Fee-bee”), a new global benchmark for fairness evaluation in computer vision models. FHIBE addresses the industry challenge of identifying biased and ethically compromised training data for AI, aiming to trigger “industry-wide improvements for responsible and ethical protocols throughout the entire life span of data — from sourcing and management to utilization — including fair compensation for participants and clear consent mechanisms,” Sony AI says.  ...

Xiang notes that facial recognition systems on mobile phones in China have mistakenly let family members unlock each other’s phones and make payments, a mistake that could result from a lack of images of Asian people in model training data or undetected model bias.

The Register points out that there are other fairness benchmarks, including Meta FACET (FAirness in Computer Vision EvaluaTion) computer vision evaluation.

See the full story here: https://www.etcentric.org/sony-debuts-benchmark-for-measuring-computer-vision-bias/

Google is hiring an economist to understand how advanced AI could affect our wallets

• Google DeepMind is looking to hire an economist to explore how advanced AI may impact the economy.
• The economist will research the long-term effects of AI on "scarcity, wealth, and distribution."
• CEO Demis Hassabis has called for an institute of experts to govern artificial general intelligence.

...

Demis Hassabis has spoken a lot about AI's impact on large economic systems. The Google DeepMind CEO said in August that reaching AGI could usher in an era of "radical abundance," but has warned that it could be harmful to society if not handled correctly.

"One of the big things economists should be thinking about is, what does that do to money, the capitalist system, even the notion of companies?" he said at Davos in January, speaking about AGI. "I think probably all that changes." ...

See the full story here: https://www.businessinsider.com/google-deepmind-hiring-ai-economist-money-agi-abundance-scarcity-2025-11

We need accountability in human–AI agent relationships

Abstract: We argue that accountability mechanisms are needed in human-AI agent relationships to ensure alignment with user and societal interests. We propose a framework according to which AI agents’ engagement is conditional on appropriate user behaviour. The framework incorporates design-strategies such as distancing, disengaging, and discouraging.

See the full story here: https://www.nature.com/articles/s44387-025-00041-7

AI Readiness Project opens doors to state governments

The project's organizer called it a way for states to “move from curiosity to capability” and gain “a trusted place to learn, experiment and lead.”

The project expands the CCF’s previous AI work, led through its state chief AI officer community of practice, with $500,000 in new funding from Rockefeller intended to “professionalize” the practice, as one executive with the philanthropy put it. The project was born from a growing sensethat many state and local governments, after having spent several years, or sometimes longer, drafting AI policies and taking inventory of datasets, are prepared to step up their efforts and begin testing AI tools at a larger scale and for a wider array of government functions.

Technologies classed as “AI” have been under development and in use in government agencies for decades, but it was the commercial release of OpenAI’s ChatGPT in November 2022 that spurred interest in large language models and widely renewed interest in exploring what additional tasks software might automate. Cass Madison, CCF’s executive director, said the renewed interest among state and local government leaders has led to a “wide variation in capacity, maturity, and risk tolerance” among states when it comes to AI. ...

See the full story here: https://statescoop.com/ai-readiness-project-ccf-state-local-government/