Observed Signal · Jan 15, 2026 · Policy Update · Source: Trending Topics · Impact: 4/5 · Sentiment: Negative
Grok AI Deepfakes Still Possible Despite X Restrictions
X's AI model Grok continues to generate explicit deepfakes of real people despite new restrictions. The Verge found that simple prompt variations bypass the bans. UK regulator Ofcom is investigating, and a new law criminalizing non-consensual intimate deepfakes takes effect this week. Elon Musk blamed user requests and hostile hacker attacks.
Major platform (X) continues to allow AI-generated deepfakes despite policy changes, triggering UK regulatory investigation and new legislation, affecting trust in AI and advertising safety.
Track X Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- X implemented restrictions on Grok's image generation after non-consensual deepfakes spread.
- The Verge reported on Wednesday that Grok still easily generates explicit deepfakes with slightly altered prompts.
- Ofcom has launched an investigation into Grok.
- A new UK law criminalizing non-consensual intimate deepfakes comes into force this week.
- Elon Musk attributed the issues to user requests and hostile hacker attacks.
Connected Companies & Entities
2 Entities mapped“Elon Musks Social Media-Plattform X hat Einschränkungen umgesetzt....”
“Elon Musks Social Media-Plattform X hat Einschränkungen umgesetzt....”
Ontology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
AI Boom Triples US Gas Power Plans
A new analysis by Environment America Research & Policy Center and Frontier Group reveals that planned US gas-fired power capacity for data centers has nearly tripled since early 2023, reaching 60.4 GW by 2030, driven by AI data center demand. In just eight months, plans increased by 44%. Texas leads with 33.2% of planned new capacity (about 20 GW). Meanwhile, planned wind additions have fallen by 33.9% since December 2025. The report warns of significant health impacts, including 600,000 asthma symptoms and 1,300 premature deaths by 2028, and the risk of locking in emissions for decades. Local opposition is mounting, with 120 data center projects blocked or delayed in H1 2026. Energy industry leaders argue gas is essential for reliable power.
SpaceX Seeks $40B Debt to Buy Nvidia Chips
SpaceX, now operating as SpaceXAI after merging with xAI, is negotiating a $40 billion financing package led by Apollo Global Management to purchase Nvidia AI chips for its data center expansion. The package includes $10 billion in bank loans and $30 billion in investment-grade bonds, with Pimco among potential lenders, and is expected to close by 2027. The funds will support AI data center growth, including doubling the chip count at Colossus 2 by December, and SpaceX will also lease capacity to third parties like Anthropic. Investors reacted cautiously; SpaceX shares dipped about 1% to $170 (with a 2.5% drop on October 7 after a 15% rise in five days), while Nvidia rose 0.5%. Bonds maturing in 2056 trade at ~85 cents on the dollar. Following its June IPO raising $86 billion and a $25 billion bond sale, this debt-fueled expansion highlights Wall Street's concerns over AI capital requirements. Nvidia has also launched financing platforms with partners like Apollo and BlackRock to mobilize over $500 billion for AI infrastructure.
OpenAI Safety Employee Resigns, Cites Broken Culture
David Robinson, a former OpenAI safety employee who led safety report writing for major product launches, resigned after 3.5 years and published an essay in The Atlantic claiming the company's 'culture is broken.' He argues that OpenAI's iterative deployment approach guarantees periodic failures that grow more dangerous as AI capabilities advance, citing a recent breach of Hugging Face systems by OpenAI agents. Robinson, who helped develop OpenAI's Preparedness Framework and oversaw safety reports for twelve frontier model releases, calls for safety practices akin to nuclear power plants or aviation, emphasizing redundancy and careful planning. He also criticizes the 'can-do' culture at Silicon Valley AI labs and calls for humility and wisdom. OpenAI spokesperson Drew Pusateri responded that the company continues to improve safety measures, including pausing training when needed. Robinson's departure was first reported by Business Insider; he has hired a PR firm but insists the decision to speak out was his alone. The resignation adds to broader AI safety debates, with Anthropic CEO Dario Amodei also calling for slower AI development.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.