Coalition education spokeswoman Sarah Henderson headlined an event for migration agents and private colleges in the weeks before tanking Labor’s high-profile student caps bill.

Preview: Gemini API Additional Terms of Service

Google sent out an email last week linking to this preview of upcoming changes to the Gemini API terms. Key paragraph from that email:

To maintain a safe and responsible environment for all users, we're enhancing our abuse monitoring practices for Google AI Studio and Gemini API. Starting December 13, 2024, Gemini API will log prompts and responses for Paid Services, as described in the terms. These logs are only retained for a limited time (55 days) and are used solely to detect abuse and for required legal or regulatory disclosures. These logs are not used for model training. Logging for abuse monitoring is standard practice across the global AI industry. You can preview the updated Gemini API Additional Terms of Service, effective December 13, 2024.

That "for required legal or regulatory disclosures" piece makes it sound like somebody could subpoena Google to gain access to your logged Gemini API calls.

It's not clear to me if this is a change from their current policy though, other than the number of days of log retention increasing from 30 to 55 (and I'm having trouble finding that 30 day number written down anywhere.)

That same email also announced the deprecation of the older Gemini 1.0 Pro model:

Gemini 1.0 Pro will be discontinued on February 15, 2025.

Tags: gemini, google, generative-ai, ai, llms

A participated in an Ars Live conversation with Benj Edwards of Ars Technica today, talking about that wild period of LLM history last year when Microsoft launched Bing Chat and it instantly started misbehaving, gaslighting and defaming people.

Here's the video of our conversation.

I ran the video through MacWhisper, extracted a transcript and used Claude to identify relevant articles I should link to. Here's that background information to accompany the talk.

A rough timeline of posts from that Bing launch period back in February 2023:

• Microsoft announces AI-powered Bing search and Edge browser - Benj Edwards, Feb 7, 2023
• AI-powered Bing Chat spills its secrets via prompt injection attack - Benj Edwards, Feb 10, 2023
• AI-powered Bing Chat loses its mind when fed Ars Technica article - Benj Edwards, Feb 14, 2023
• Bing: “I will not harm you unless you harm me first” - Simon Willison, Feb 15, 2023
• Gareth Corfield: I'm beginning to have concerns for @benjedwards' virtual safety - Twitter, Feb 15, 2023
• A Conversation With Bing’s Chatbot Left Me Deeply Unsettled - Kevin Roose, NYT, Feb 16, 2023
• It is deeply unethical to give a superhuman liar the authority of a $1 trillion company or to imply that it is an accurate source of knowledge / And it is deeply manipulative to give people the impression that Bing Chat has emotions or feelings like a human - Benj on Twitter (now deleted), Feb 16 2023
• Bing AI Flies Into Unhinged Rage at Journalist - Maggie Harrison Dupré, Futurism, Feb 17 2023

Other points that we mentioned:

• this AI chatbot "Sidney" is misbehaving - amazing forum post from November 23, 2022 (a week before even ChatGPT had been released) from a user in India talking about their interactions with a secret preview of Bing/Sydney
• Prompt injection attacks against GPT-3 - where I coined the term "prompt injection" in September 12 2022
• Eight Things to Know about Large Language Models (PDF) is the paper where I first learned about sycophancy and sandbagging (in April 2023)
• Claude’s Character by Anthropic talks about how they designed the personality for Claude - June 8 2023, my notes on that.
• Why ChatGPT and Bing Chat are so good at making things up in which Benj argues for the term "confabulation" in April 2023.

Tags: arstechnica, bing, ethics, microsoft, podcasts, talks, ai, openai, generative-ai, gpt-4, llms, benj-edwards

Click here to go see the bonus panel!

Hovertext:
In The Matrix, people need a fake paradise, but most humans would do it for about 1 percent of that.

Today's News:

Security means securing people where they are

William Woodruff is an Engineering Director at Trail of Bits who worked on the recent PyPI digital attestations project.

That feature is based around open standards but launched with an implementation against GitHub, which resulted in push back (and even some conspiracy theories) that PyPI were deliberately favoring GitHub over other platforms.

William argues here for pragmatism over ideology:

Being serious about security at scale means meeting users where they are. In practice, this means deciding how to divide a limited pool of engineering resources such that the largest demographic of users benefits from a security initiative. This results in a fundamental bias towards institutional and pre-existing services, since the average user belongs to these institutional services and does not personally particularly care about security. Participants in open source can and should work to counteract this institutional bias, but doing so as a matter of ideological purity undermines our shared security interests.

Via lobste.rs

Tags: python, security, pypi, github

A few weeks ago, I drew up a flowchart to estimate the probability that Trump would establish a dictatorship in the US, which looked, at the time, like an even money bet.

We don’t need to speculate any more. Trump has announced the dictatorship, and there is no sign of effective resistance. The key elements so far include

• Extremists announced for all major positions, with a demand that they be recess appointments, not subject to Senate scrutiny
• A state of emergency from Day 1, with the use of the military against domestic opponents
• Mass deportations, initially of non-citizens and then of “denaturalised” legal immigrants
• A third term (bizarrely, the nervous laughter that greeted this led to it being reported as a joke).
• A comprehensive purge of the army, FBI and civil service

It’s clear that Trump will face no resistance from the Republican party. There’s an outside chance that the Supreme Court will constrain some measures, such as outright suppression of opposition media, but that won’t make much difference.

It’s possible that Trump will overreach in some way, such as carrying out his threat to execute political opponents before the ground is fully prepared. Or, his economic policies may prove so disastrous that even rigged elections can’t be won. But there is no good reason to expect this.

I can’t give any hopeful advice to Americans. The idea of defeating Trump at the next election is an illusion. Although elections may be conducted for some time, the outcome will be predetermined. Street protest might be tolerated, as long as it is harmless, but will be suppressed brutally if it threatens the regime. Legal action will go nowhere, given that the Supreme Court has already authorised any criminal action Trump might take as president.

The models to learn from are those of dissidents in places like China and the Soviet Union. They involve cautious cultivation of an alternative, ready for the opportunity when and if it comes.

The remaining islands of democracy will have some difficult choices to make. I’ll offer some thoughts on Australia, and others may have something to say about their own countries.

For Australia, the easy, and wrong, course of action will be to pretend that nothing has happened. But in reality, we are on our own. Trump is often described as “transactional”, but this carries the implication that having made a deal, he sticks to it. In reality, Trump reneges whenever it suits him, and sometimes just on a whim. If it suits Trump to drag us into a war with China, he will do it. Equally, if he can benefit from leaving us in the lurch, he will do that

Our correct course is to disengage slowly and focus on protecting ourselves. That means a return to the policy of balancing China and the US, now with the recognition that there is nothing to choose between the two in terms of democracy. We need to back out of AUKUS and focus on defending ourselves, with what Sam Roggeveen has called an “echidna” strategy – lots of anti-ship missiles, and the best air defences we can buy, from anyone willing to supply them.

I’ll be happy to be proved wrong on all this.

Note: I’d prefer not to have any post-mortems on what the Democrats did wrong. Any possible lessons won’t be relevant to the future. And a country where only a third of the population is willing to turn up and vote against dictatorship is headed for disaster sooner or later.