Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1B in damages

https://www.tomshardware.com/tech-industry/artificial-intelligence/anthropics-proposed-ai-safety-bill-only-kicks-in-if-50-deaths-or-usd1-billion-in-damages-are-on-the-line-proposal-would-effectively-make-the-commerce-department-the-only-agency-capable-of-stopping-an-ai-model

32•f-serif•46m ago

Comments

serioussecurity•24m ago

$1B is a pretty easy bar to clear.

kfichter•20m ago

$1B is an extraordinarily easy bar to clear with a single crypto hack and Anthropic was repeatedly refusing to give access to Mythos to defenders in the crypto world. I think this stems from bias rather than a conspiracy theory but it's an entirely avoidable disaster waiting to happen.

LoganDark•12m ago

> Anthropic was repeatedly refusing to give access to Mythos to defenders in the crypto world.

This almost makes sense. You almost can't risk vulnerabilities even being discovered in certain systems that are impossible to patch, like immutable distributed smart contracts.

Not sure that justifies withholding Mythos access entirely, but what do I know. I got access to Mythos a few days ago apparently since Anthropic folded Project Glasswing into the CVP.

mrtksn•20m ago

On the bright side, don’t offer an exchange rate based on ethnicity or other background criteria but I get the feeling that your AI won’t be able to kill 49 billionaires without triggering something. Can probably do 49 millionaires without hitting the $1B threshold? Maybe someone will come up with a risk management SaaS as not all millionaires are the same and maybe you’ll hit the $1B before the 49th.

LoganDark•14m ago

I feel like the surviving family of even a single millionaire could claim $1B in damages if the millionaire was young enough.

mrtksn•7m ago

Right, so maybe they had an exchange rate after all. Just implicit.

I got the feeling that what they had in their essentially culling the energy consumers with low economic output.

LoganDark•6m ago

There's always been an exchange rate based on the quality of your legal team. Get one sociopathic enough and they'll pick a judge right apart.

airstrike•19m ago

The next step is demanding anyone who wishes to release a Certified Safe AI™ model to prove they can insure the $1B in damages and pass a compliance audit for how they will prevent the LTM death count from reaching 50, effectively making it so that only big labs can serve government and enterprise.

Muromec•7m ago

With existing players given a free pass as is necessary for the functioning of a democratic society

verdverm•2m ago

There is an idea for a pooled, mutualized insurance fund. Ai companies would contribute based on various factors

https://www.lawfaremedia.org/article/bootstrapping-frontier-...

karanbhangui•19m ago

so 49 totally okay.

LoganDark•16m ago

I would not like to see a world where companies push their models as close as possible to killing exactly 49 people without reaching 50

mikeodds•18m ago

I guess they solved the trolley problem internally

LoganDark•17m ago

Do we even know if ChatGPT's sycophancy killed 50 people? The last figure I've seen for suicide-by-ChatGPT was like 13 or 14 cases, not sure if there have been more.

verdverm•16m ago

This alleviates them from harm on the individual level, consumers and communities, we the people. We've already seen encouraged suicides and school shootings. These would not qualify.

That is why the bar is set where it is. The corporate circle jerk designed this as a cost of doing business, not as a check on responsibility or restitution for real damages to people and society.

1228z•12m ago

After the AISI and MAISI AI promotion institutes, let us create SISI (pronounced sissy) for investigating a super-intelligence safety bill.

This is all theater, The whole Trump clan is invested in the AI bubble.

pluc•9m ago

Who's keeping score? Also define "death". If a non-alive person has social media, a blog, a diary, meeting recordings, emails - we'll make it so they never left. Hell, we'll partner with Elon and Mark and a make it a deactivation hook.

pfdietz•7m ago

Seems kind of low, actually.

giancarlostoro•6m ago

Not sure what kicks in, but if a tiny company worth very little is destroyed by an AI hack they should be 100% liable for all damages.

Why should a small business suffer because it wasnt worth a billion dollars?

10xDev•3m ago

So Claude figured out a human life is worth 20 million dollars?

Show HN: BricksChat – Generate Lego builds from images, including Technic

Omarchy Gets the Web Browser Launching Faster by Avoiding XDG Overhead

Negotiation tips for people who hate negotiating

Why we decided to open source Relate

Aligning a Personal Site with Indieweb and Web Standards

GhostOS – Lightweight, Offline Desktop AI Assistant (macOS/Win/Linux)

Gnome OS Preview

A house on the internet that feels like home

Metas smart glasses record conversations without the wake phrase, company admits

Will it be a golden age or a dark age for mathematics (and science)?

Websites are closing their doors to AI agents

Visual test failed. Is it the page or the capture?

ChatGPT Dot Review: Your Chief of Staff Is Still an Intern

Staying off the hot path

RentAHuman

How a Computer System Works (1975 book)

I tried the vibe coded Fusion 360 (it's disappointing) [video]

Show HN: ElNeo, a E2E enabled matrix client for Emacs, inspired by telega

The Law of the Instrument

Nordwand Mono – A monospace font for coders and adventurers

Java sealed classes and exhaustive pattern matching

Why big dogs age faster

Show HN: TaskPocket, a to-do app with an MCP server instead of built-in AI

World Bank says 'small AI' could transform developing countries

Nvidia Nemotron 3.5 Super VL Powers Specialized Agents with Multimodal Reasoning

A Note on Content Restrictions

Italy's guiltiest food ritual went public

Nerd Speak

High performance policy routed IPsec mesh SD-WAN

Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1B in damages

Show HN: BricksChat – Generate Lego builds from images, including Technic

Omarchy Gets the Web Browser Launching Faster by Avoiding XDG Overhead

Negotiation tips for people who hate negotiating

Why we decided to open source Relate

Aligning a Personal Site with Indieweb and Web Standards

GhostOS – Lightweight, Offline Desktop AI Assistant (macOS/Win/Linux)

Gnome OS Preview

A house on the internet that feels like home

Metas smart glasses record conversations without the wake phrase, company admits

Will it be a golden age or a dark age for mathematics (and science)?

Websites are closing their doors to AI agents

Visual test failed. Is it the page or the capture?

ChatGPT Dot Review: Your Chief of Staff Is Still an Intern

Staying off the hot path

RentAHuman

How a Computer System Works (1975 book)

I tried the vibe coded Fusion 360 (it's disappointing) [video]

Show HN: ElNeo, a E2E enabled matrix client for Emacs, inspired by telega

The Law of the Instrument

Nordwand Mono – A monospace font for coders and adventurers

Java sealed classes and exhaustive pattern matching

Why big dogs age faster

Show HN: TaskPocket, a to-do app with an MCP server instead of built-in AI

World Bank says 'small AI' could transform developing countries

Nvidia Nemotron 3.5 Super VL Powers Specialized Agents with Multimodal Reasoning

A Note on Content Restrictions

Italy's guiltiest food ritual went public

Nerd Speak

High performance policy routed IPsec mesh SD-WAN

Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1B in damages

Anthropic's proposed AI safety bill only kicks in if 50 deaths or $1B in damages

Comments