Lightnews — Scholar-powered news

Reposted by Amos Toh

Spencer Reynolds @sreynolds.bsky.social · 22d

NEW: Internal report reveals DHS intelligence portal was accidentally opened to tens of thousands of users for months, exposing sensitive intel about Americans. @brennancenter.org obtained the docs. @wired.com has the story. 🧵

www.wired.com/story/a-dhs-...

www.wired.com

36 370 650

Reposted by Amos Toh

Joseph Nunn @josephanunn.bsky.social · Aug 11

THREAD: This morning, President Trump and members of his cabinet announced a set of sweeping, unprecedented, and unwarranted actions to impose federal control over local policing in Washington, D.C.

2 23 41

Reposted by Amos Toh

Margy O’Herron @margyoh.bsky.social · Aug 7

@amostoh.bsky.social explains what the new AI EO is really about:
“These restrictions feign a commitment to truth and impartiality while expecting tech companies to fall in line with the administration’s version of reality.”

Amos Toh @amostoh.bsky.social · Aug 4

A more comprehensive analysis for @brennancenter.org of how the Trump administration's latest AI policy could leave us more divided and less informed: www.brennancenter.org/our-work/ana...

3 7

Amos Toh @amostoh.bsky.social · Aug 6

According to my @brennancenter.org colleague @sreynolds.bsky.social, DHS' secretive intelligence arm "analyzes contents of phones or laptops seized by agencies like CBP and ICE, comparing them with personal data to create detailed pictures of people, their social networks, and their habits"

The American Prospect @prospect.org · Aug 6

Drones, spy blimps, and motion-triggered cameras—the border is now a testing ground for dystopian surveillance. From James Baratta, the high-tech machine behind the border industrial complex:
trib.al/xNL3uaj

Surveillance at the Border

It’s only a matter of time before the drones, spy blimps, license plate readers, and motion-activated cameras come to the rest of America.

trib.al

1 4

Reposted by Amos Toh

Spencer Reynolds @sreynolds.bsky.social · Aug 6

Essential reading on border surveillance tech, its human impact, and the risk it will move inward.

As part of this apparatus, a U.S. Intelligence Community element helps CBP analyze the phones and laptops it seizes at the border. I spoke with the author about the activity's lax rules and risks.

The American Prospect @prospect.org · Aug 6

Drones, spy blimps, and motion-triggered cameras—the border is now a testing ground for dystopian surveillance. From James Baratta, the high-tech machine behind the border industrial complex:
trib.al/xNL3uaj

Surveillance at the Border

It’s only a matter of time before the drones, spy blimps, license plate readers, and motion-activated cameras come to the rest of America.

trib.al

2 4

Amos Toh @amostoh.bsky.social · Aug 4

A more comprehensive analysis for @brennancenter.org of how the Trump administration's latest AI policy could leave us more divided and less informed: www.brennancenter.org/our-work/ana...

16 30

Amos Toh @amostoh.bsky.social · Jul 24

25/ The effort to purge “woke AI” might force the hand of tech companies, leading to AI that is even more broken. This could saddle the military with faulty intel that harms soldiers and civilians. For the rest of us, we would be stuck with technology that leaves us more divided and less informed.

1

Amos Toh @amostoh.bsky.social · Jul 24

24/ OpenAI and Anthropic are selling their foundation models to the military.

Google is working with Lockheed to develop AI weapons.

Meta has bought a 49% stake in Scale AI, an up-and-coming defense contractor.

Amazon and Microsoft are major military cloud providers.

Billions are at stake.

1 1

Amos Toh @amostoh.bsky.social · Jul 24

23/ This time, tech company leaders have tried to pander to the administration’s “anti-woke” agenda, but this strategy could put them in a bind, particularly as they seek to land lucrative defense contracts that would offset the enormous costs of their AI business.

fortune.com/2025/01/13/z...

Zuckerberg says most companies need more ‘masculine energy’

Zuckerberg, who launched his career by rating the attractiveness of women at Harvard University, lamented the rise of “culturally neutered” companies.

fortune.com

1

Amos Toh @amostoh.bsky.social · Jul 24

22/ Amazon complained this was retaliation against Jeff Bezos – Trump was unhappy at the time about reporting by WaPo, which Bezos owns. The suit was later dismissed after the contract was revised, and Amazon was awarded a portion of it.

www.reuters.com/legal/govern...

U.S. judge ends Amazon challenge to $10 bln cloud contract after Pentagon cancellation

A U.S. judge on Friday dismissed Amazon.com's legal challenge to the Defense Department's2019 decision to award a $10 billion JEDI cloud-computing project to rival Microsoft Corp (MSFT.O) after the Pentagon canceled the contract.

www.reuters.com

1

Amos Toh @amostoh.bsky.social · Jul 24

21/ This wouldn’t be the first time the Trump admin has tried to kill federal contracts in response to perceived slights. In 2019, Amazon sued the first Trump admin for using “improper pressure” to deprive it of a $10 billion military cloud computing contract.

www.nytimes.com/2019/12/09/t...

Amazon Accuses Trump of ‘Improper Pressure’ on JEDI Contract (Published 2019)

www.nytimes.com

1

Amos Toh @amostoh.bsky.social · Jul 24

20/ But the wild card here is the administration itself – a chatbot response goes viral, Trump takes offense, and threatens to cancel the provider’s federal contracts – sounds familiar?

www.wsj.com/business/tru...

Exclusive | Trump Aides Discussed Ending Some SpaceX Contracts, but Found Most Were Vital

Fallout between the president and the rocket maker’s billionaire founder threatened the company’s multibillion-dollar agreements with the government

www.wsj.com

1

Amos Toh @amostoh.bsky.social · Jul 24

19/ The order provides room for this, clarifying that agencies should “account for technical limitations” in enforcing compliance and “avoid over-prescription and afford latitude for vendors.”

1

Amos Toh @amostoh.bsky.social · Jul 24

18/ Of course, tech companies could treat the order as symbolic, and placate the admin with messaging about anti-woke measures rather than meaningful changes to how their models are developed and used.

1

Amos Toh @amostoh.bsky.social · Jul 24

17/ The White House guidance does not apply to national security agencies, and the order allows for exceptions “as appropriate” for nat sec uses of Large Language Models. But the intent to mold foundation models in the admin’s image of what “truths” are acceptable is clear.

1

Amos Toh @amostoh.bsky.social · Jul 24

16/ This risk is not theoretical: in April, the White House updated its guidance on federal use and acquisition of AI to remove all references to bias detection and mitigation, including measures to account for the tech’s impact on underrepresented populations. www.lawfaremedia.org/article/narr...

Narrowing the National Security Exception to Federal AI Guardrails

Fostering public trust in how the government uses AI to protect national security requires robust and enforceable rules on how it is authorized, tested, disclosed, and overseen.

www.lawfaremedia.org

1

Amos Toh @amostoh.bsky.social · Jul 24

15/ While the order is limited to Large Language Models, it’s easy to see how defense contractors could be pressured to strip out bias detection the admin doesn't like across foundation model training. This may undermine other military functions that increasingly rely on AI, like target recognition.

1

Amos Toh @amostoh.bsky.social · Jul 24

14/ If federal providers of AI-based translation services are prevented from fine tuning their models to be more accurate for languages other than in English, or to limit the reinforcement of harmful stereotypes, this could compromise the analysis of intelligence in these languages.

1 1

Amos Toh @amostoh.bsky.social · Jul 24

13/ The problem is that this could harm model performance. And the federal agency that this would impact the most is the government’s biggest spender on AI - the Department of Defense. www.brookings.edu/articles/the...

The evolution of artificial intelligence (AI) spending by the U.S. government | Brookings

U.S. federal government spending on AI-related contracts has massively increased in the last year, especially in the defense sector.

www.brookings.edu

1

Amos Toh @amostoh.bsky.social · Jul 24

12/ Tech companies may also try to comply by rolling back internal measures to mitigate the spread of content that denigrates or erases racial and other minorities, or account for underrepresented languages in training data.

Critical studies like this could be shelved: openai.com/index/evalua...

Evaluating fairness in ChatGPT

We've analyzed how ChatGPT responds to users based on their name, using language model research assistants to protect privacy.

openai.com

1

Amos Toh @amostoh.bsky.social · Jul 24

11/ But fine-tuning is more insidious than content filters, since it biases the model to favor certain outcomes. We could end up with models that systematically produce biased responses in a neutral-sounding way, e.g. that there is no scientific consensus on the effects of climate change.

1

Amos Toh @amostoh.bsky.social · Jul 24

10/ This type of “fine-tuning” is apparently used to prevent chatbots from teaching users how to build bombs, conduct cyberattacks, or engage in dangerous or illegal action. Again, not foolproof - users can manipulate prompts to evade these restrictions.

www.theguardian.com/technology/a...

AI chatbots’ safeguards can be easily bypassed, say UK researchers

Five systems tested were found to be ‘highly vulnerable’ to attempts to elicit harmful responses

www.theguardian.com

1

Amos Toh @amostoh.bsky.social · Jul 24

9/ Another way to comply would be to nudge Large Language Models towards certain types of answers i.e. getting the model to generate answers deemed “untruthful” or “biased,” before telling it that alternative responses are more accurate. In theory, this would steer models towards the latter.

1

Amos Toh @amostoh.bsky.social · Jul 24

8/ It also sets up an impossible line drawing exercise for tech companies: if you can’t query a chatbot about climate change, what about global warming? Or greenhouse gases? “Climate harms”?

1

Amos Toh @amostoh.bsky.social · Jul 24

7/ These filters are imprecise and crude - asking about Jonathan *L.* Zittrain is enough to elicit a response. With this kind of slippage, it’s unclear if using content filters will satisfy the admin, while making the models more frustrating to use.

1