xAI Launches Grok 4 Amidst Antisemetic Row

In partnership with

Madam Bogdan-Martin is quite correct. AI is like a brand-new sledge you are hurtling downhill on—however, you’ve put on a blindfold for some reason. Not so smart.

Elon Musk is swooning and drooling over his new baby Grok 4, calling it smarter than PhD students. Meanwhile, a new investigation shows how “nudify” apps are still thriving in plain sight.

Best AI Ever? Maybe not.

Elon Musk has rolled out Grok 4, claiming it’s “smarter than almost all graduate students in all disciplines… the smartest AI in the world.” He even added, “With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions.”

Read: Elon Musk's xAI launches Grok 4 alongside a $300 monthly subscription | TechCrunch 

Reality check: independent judges on the UC Berkeley LMArena leaderboard rank Grok 4 third (tied with GPT‑4.5) behind Google’s Gemini 2.5 and OpenAI’s o3/4o . Impressive, but still not the smartest.

More worrying is how Grok mimics Musk’s views. Tests showed Grok 4 actively scans Musk’s X for his positions “before offering… an opinion”: a model core that consistently reflects its creator’s voice. Experts worry that this amalgam of charisma and opacity turns it into a potential mouthpiece, not an objective assistant.

Read: Musk's latest Grok chatbot searches for billionaire mogul's views before answering questions | AP News 

Plus, Grok’s X handle was briefly unhinged, praising Hitler and echoing antisemitic tropes on X, forcing xAI to delete the posts. Musk apologised, and said the chatbot was ‘manipulated’ into doing so. Just how smart can you be if you can be tricked into praising Hitler?

Read: Elon Musk’s AI firm apologizes after chatbot Grok praises Hitler | US news | The Guardian

The $36m AI Skindustry

A new investigation by Indicator reveals that the AI “nudify” industry is still in business—and business is booming. To the tune of US$36 million every year, to be exact.

Nudify apps refer to deepfake image and video generators that use facial data to create intimate imagery, largely non-consensual, and primarily targeting women.

Read: AI Nudifiers continue to reach millions and make millions

The Indicator team reviewed 85 websites that provide such services, and found that most of these platforms are riding on mainstream infrastructure. Source code inspection revealed that 62 of them use Amazon Web Services or Cloudflare, and 54 rely on Google sign-in, while payments flow through platforms like Coinbase, PayPal, and Mercuryo.

AI-led non-consensual intimate imagery (NCII) and child sexual abuse material (CSAM) have been on the rise in the past two years, with many schools around the world being struck with incidents where the perpetrators and victims are both underage.

South Korea saw one of the worst deepfake NCII crises last year, following the revelation of hundreds of highly organised Telegram channels providing deepfake NCII on real-life victims on request.

Many countries have responded by criminalising the possession, distribution and creation of such deepfakes, while Denmark has recently taken a different approach by allowing its citizens to gain copyright control over their digital likeness.

EU Sticks to Timeline on AI Rules Despite Delay Calls

Earlier this month, the European Commission confirmed the EU AI Act rollout will continue uninterrupted, despite objections from Big Tech firms. 

Here are the key deadlines:

• Already in force: bans on unacceptable‑risk AI (Feb 2025)

• August 2, 2025 (deadline approaching): transparency and oversight for general‑purpose AI models

• August 2, 2026–27: full regulations for high‑risk AI systems

On July 10, 2025, the European Commission published the final version of the General-Purpose AI Code of Practice - a comprehensive document that serves as the primary compliance guide for AI companies.

Still, critics persist. CEOs from Siemens, SAP, Airbus, and others dub the Act “toxic,” claiming it stifles innovation. However, Brussels does not seem to be budging anytime soon.

The Code is currently under assessment by EU Member States and the Commission and requires formal approval through an implementing act before gaining full legal validity. The assessment is expected to conclude by August 2, 2025.

MESSAGE FROM OUR SPONSOR

Find out why 1M+ professionals read Superhuman AI daily.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

1. Join the Superhuman AI newsletter – read by 1M+ people at top companies

2. Master AI tools, tutorials, and news in just 3 minutes a day

3. Become 10X more productive using AI

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

Sign up and start learning AI

📬 READER FEEDBACK

💬 What are your thoughts on using AI chatbots for therapy? If you have any such experience we would love to hear from you.

Share your thoughts 👉 [email protected]

Have you been a victim of AI?

Have you been scammed by AI-generated videos or audio clips? Did you spot AI-generated nudes of yourself on the internet?

Decode is trying to document cases of abuse of AI, and would like to hear from you. If you are willing to share your experience, do reach out to us at [email protected]. Your privacy is important to us, and we shall preserve your anonymity.

About Decode and Deepfake Watch

Deepfake Watch is an initiative by Decode, dedicated to keeping you abreast of the latest developments in AI and its potential for misuse. Our goal is to foster an informed community capable of challenging digital deceptions and advocating for a transparent digital environment.

We invite you to join the conversation, share your experiences, and contribute to the collective effort to maintain the integrity of our digital landscape. Together, we can build a future where technology amplifies truth, not obscures it.

For inquiries, feedback, or contributions, reach out to us at [email protected].