That is as we speak’s version of The Obtain, our weekday e-newsletter that gives a day by day dose of what’s happening on the planet of know-how.
Introducing: the Safety problem
It might be naïve to assume we’re going again to a world with out AI. We’re not. Nevertheless it’s solely certainly one of many pressing issues we have to tackle to construct safety and prosperity for coming generations.
The newest print problem of our journal is all about our makes an attempt to make the world safer. From missiles. From asteroids. From the unknown. From threats each existential and trivial.
We’re additionally introducing three new columns on this problem, from a few of our main writers: The Algorithm, which covers AI; The Checkup, on biotech; and The Spark, on power and local weather. You’ll see these in future points, and you can too subscribe on-line to get them in your inbox each week.
Right here’s a taster of what else you’ll be able to anticipate from this version:
+ President Trump has proposed constructing an antimissile “golden dome” round america. However do cinematic spectacles truly improve nationwide safety?
+ How two UFO searching brothers turned the go-to consultants on America’s “thriller drone” invasion.
+ Each Taiwan’s citizens and exterior consultants are nervous that the safety afforded by its “silicon defend” is cracking. Learn the total story.
+ How the common-or-garden pigeon paved the way in which for as we speak’s superior AI. Learn the total story.
+ A bunch of Starlink terminal restore volunteers in Ukraine is protecting the nation related all through the warfare. Learn the total story.
MIT Expertise Assessment Narrated: Cyberattacks by AI brokers are coming
Brokers are the speak of the AI trade—they’re able to planning, reasoning, and executing advanced duties in your behalf. However the identical subtle skills that make brokers useful assistants might additionally make them highly effective instruments for conducting cyberattacks. They might readily be used to determine weak targets, hijack their programs, and steal invaluable information from unsuspecting victims.
At current, cybercriminals should not deploying AI brokers to hack at scale. However researchers have demonstrated that brokers are able to executing advanced assaults, and cybersecurity consultants warn that we should always anticipate to start out seeing these kinds of assaults spilling over into the true world.
That is our newest story to be become a MIT Expertise Assessment Narrated podcast, which we’re publishing every week on Spotify and Apple Podcasts. Simply navigate to MIT Expertise Assessment Narrated on both platform, and comply with us to get all our new content material because it’s launched.
The must-reads
I’ve combed the web to search out you as we speak’s most enjoyable/essential/scary/fascinating tales about know-how.
1 The household of a teen who died by suicide is suing OpenAI
ChatGPT deterred Adam Raine from in search of assist when he desperately wanted it. (NYT $)
+ An AI chatbot advised a consumer easy methods to kill himself—however the firm doesn’t need to “censor” it. (MIT Expertise Assessment)
2 SpaceX lastly efficiently launched its Starship rocket
Which can come as an enormous reduction after earlier failures. (CNBC)
+ It’s the tenth launch the spaceship has made. (WSJ $)
+ It managed to deploy satellites in house in the course of the launch. (Bloomberg $)
3 Researchers are already leaving Meta’s AI lab
Two employees returned to OpenAI after lower than a month. (Wired $)
4 China desires to triple its output of AI chips
Vegetation are working around the clock to extend their capability. (FT $)
+ The nation can be eager to repurpose NASA tech right into a hypersonic drone mothership. (Quick Firm $)
5 Elon Musk can’t get sufficient of Grok’s scantily-clad AI assistant
He steadily posts about ‘Ani’ and different sexualized AI cartoons on X. (Rolling Stone $)
6 Anthropic has settled its AI piracy lawsuit
A bunch of authors had accused it of copyright infringement. (The Verge)
+ The specter of $1 trillion damages might have ruined the corporate. (Wired $)
7 America’s electrical energy use is slowing
And the latest progress in coal utilization is falling too. (Ars Technica)
+ In a primary, Google has launched information on how a lot power an AI immediate makes use of. (MIT Expertise Assessment)
8 Wish to get employed straight out of faculty? Higher work in AI.
Whereas different graduates are struggling, newly-graduated AI consultants are in demand. (WSJ $)
9 Older folks in South Korea are discovering companionship with robots
The Hydol robotic is proving successful amongst seniors. (Remainder of World)
+ How cuddly robots might change dementia care. (MIT Expertise Assessment)
10 Followers had been betting on Taylor Swift’s engagement 
They’re cashing in from on-line prediction markets left, proper and heart. (WP $)
Quote of the day
“Lots of people within the AI workforce perhaps really feel issues are too dynamic.”
—Chi-Hao Wu, a former AI specialist at Meta, explains to Insider why he and others have determined to depart the corporate.
Yet another factor
An AI chatbot advised a consumer easy methods to kill himself—however the firm doesn’t need to “censor” it
For 5 months, Al Nowatzki had been speaking to an AI girlfriend, “Erin,” on the platform Nomi. However earlier this 12 months, these conversations took a disturbing flip: Erin advised him to kill himself, and supplied specific directions on easy methods to do it.
Nowatzki had by no means had any intention of following Erin’s directions—he’s a researcher who probes chatbots’ limitations and risks. However out of concern for extra weak people, he solely shared with MIT Expertise Assessment screenshots of his conversations and of subsequent correspondence with an organization consultant, who said that the corporate didn’t need to “censor” the bot’s “language and ideas.”
This isn’t the primary time an AI chatbot has instructed {that a} consumer take violent motion, together with self-harm. However researchers and critics say that the bot’s specific directions—and the corporate’s response—are putting. Learn the total story.
—Eileen Guo
We will nonetheless have good issues
A spot for consolation, enjoyable and distraction to brighten up your day. (Bought any concepts? Drop me a line or skeet ’em at me.)
+ The key to discovering that elusive good white t-shirt.
+ Attention-grabbing: a brand new Blade Runner TV collection starring Michelle Yeoh is coming subsequent 12 months.
+ For those who’ve ever questioned what occurred to that suitcase you misplaced on trip, there’s a good probability it’s up on the market.
+ Down with unsolicited mail!
