The Download: The legal clash between Musk and Altman, and the profitability issue of AI

Musk, a co-founder of OpenAI, asserts that he was misled into financing the company based on false assumptions. He is pursuing $134 billion in compensation, along with the dismissal of Altman and president Greg Brockman, and a return of the organization to a non-profit status.

Discover how the trial may disrupt the global AI landscape.

—Michelle Kim

The gap between excitement and earnings

In a well-known South Park episode, a group of gnomes sneak out at night to swipe underpants. What’s their plan? The gnomes reveal their pitch presentation. “Phase 1: Acquire underpants. Phase 2: ? Phase 3: Earnings.” It’s a business strategy that reflects the current landscape of AI. 

Firms have developed the technology (Step 1) and promised transformation (Step 3). However, the pathway to that is still uncertain. Read about possible future avenues.

—Will Douglas Heaven

This article was initially published in The Algorithm, our weekly newsletter providing you insights on everything AI. Subscribe to have it delivered to your inbox every Monday.

Entering a new era of manipulated deepfakes

For years, specialists have cautioned that deepfakes could be used for harmful purposes. Those threats are now a reality.

Inexpensive, widely available models now generate weaponized deepfakes—from explicit sexual content to political misinformation—that appear alarmingly authentic. They’re already provoking violence, influencing opinions, and creating suspicion, particularly affecting women and marginalized communities.

Experts express concern that they’re eroding trust and critical reasoning. Here’s what has them worried.

—Eileen Guo

Weaponized deepfakes feature prominently on our 10 Significant Trends in AI Right Now, MIT Technology Review’s guide to what truly merits your focus in the hectic and vibrant realm of AI.

The essential reads

I’ve searched through the web to bring you today’s most entertaining/important/alarming/captivating stories regarding technology.

1 OpenAI has concluded its exclusive agreement with Microsoft
The new arrangement enables OpenAI to pursue competitors like Amazon. (Reuters $)
+ Microsoft will continue to license OpenAI’s technology, but not on an exclusive basis. (NYT $)
+ OpenAI is falling short of essential growth targets before its IPO. (WSJ $)

2 Google has inked a top-secret AI agreement with the Pentagon
It authorizes AI use for “any lawful government objective.” (The Information $)
+ More than 600 Google employees had urged a halt to the agreement. (QZ)
+ AI companies are expected to train military versions of their models with classified information. (MIT Technology Review)

3 The EU has instructed Google to make Android available to AI competitors
It aims to eliminate Gemini’s inherent advantage. (Ars Technica)
+ Google describes the action as an “unjustified intervention.” (WSJ $)
+ A final ruling is anticipated by the end of July. (Reuters $)

4 OpenAI is reportedly working on an AI-first smartphone
It would replace applications with agents. (TechCrunch)
+ Qualcomm and MediaTek may be producing its processors. (Gizmodo)

5 A brain implant aimed at depression is advancing to human trials
The FDA has authorized a human study of the device. (Wired $)
+ Brain-computer interfaces have thus far struggled to make it to market. (MIT Technology Review)

6 A populist uprising against AI is gaining traction in rural America
From Indiana to Idaho, voters are responding against the technology. (NYT $)
+ Anti-AI demonstrations are growing globally. (MIT Technology Review)

7 DeepSeek has set its new model’s price at 97% lower than OpenAI’s GPT-5.5
It aims to draw more enterprises, developers, and users dependent on agents. (SCMP)
+ Here are three reasons why DeepSeek V4 is significant. (MIT Technology Review)

8 AI is now responsible for a third of newly created websites
A study discovered it is making the web more cheerful and less wordy. (404 Media)

9 Leading talent is departing Big Tech to establish their own AI ventures
Meta, Google, and OpenAI are experiencing a talent exodus. (CNBC)

10 Taylor Swift is seeking to trademark her voice and likeness
The Grammy-winning artist has been the focus of multiple deepfakes. (NBC News)
+ An increasing number of celebrities are countering AI with trademarks. (BBC)

Quote of the day

“The truth is people aren’t fond of him.”

—Judge Yvonne Gonzalez Rogers responds to potential jurors revealing their unfavorable opinions of Elon Musk prior to his trial with Sam Altman, The Verge reports.

One More Thing

How covid conspiracy theories have sparked a troubling return of AIDS denialism

When Joe Rogan inaccurately claimed that “party drugs” were a “significant contributor to AIDS,” millions tuned in. He also asserted that AZT, the first medication used to treat AIDS, killed individuals “faster” than the disease itself—another assertion that has been debunked.

These remarks highlight an undeniable resurgence in AIDS denialism: a set of false theories claiming either that HIV doesn’t cause AIDS or that HIV doesn’t exist. By the start of the 2000s, these ideas had mostly faded. That changed with the emergence of the coronavirus.

Trace the digital journey from Covid skepticism to the revival of a dangerous conspiracy theory.

—Anna Merlan

We can still enjoy nice things

A space for comfort, enjoyment, and distraction to uplift your day. (Have any suggestions? Send me a message.)

+ Discover the planets from your computer with this live sky map.
+ This DJ set from Daphni is a remarkable expedition through electronic music.
+ NASA’s breathtaking Artemis II wallpapers deliver a high-resolution slice of deep space to your phone.
+ This intriguing GPS explanation clarifies how your phone determines your precise location.