Sitting here with a flat white gone cold, I’ve been flicking through the latest on Grok from xAI. Over the past 14 days, it’s picked up a policy clampdown on dodgy image edits and whispers of Grok 5 edging closer, with alpha access kicking off for select users this January. Nothing earth-shattering, but these nudges make it more reliable for everyday tasks, whether you’re drafting emails or checking data.
—
✅ Grok AI Policy Update Blocks Deepfake Bikini Images
On 15 January 2026, X tightened Grok’s image tools to stop non-consensual deepfakes, especially bikini-style edits of real people. It’s a response to regulator pressure and user complaints, with technical blocks now in place, though some tests show gaps remain.[7]
For marketers, this means safer campaign visuals without risking backlash from altered celeb shots. Developers can build apps with less worry about misuse liability, like generating product mockups tied to Shopify listings. Content writers get a cleaner tool for blog thumbnails, pulling from real briefs without ethical slip-ups.
—
✅ Grok 5 Alpha Access Rolls Out in January 2026
Rumours solidified into plans: Grok 5 alpha testing starts this January for X Premium+ subscribers and enterprise partners, focusing on agentic features ahead of public beta in February.[3]
Analysts can auto-summarise call transcripts from long Zoom recordings, spotting trends faster. Developers sync inventory queries with real-time X data for apps. Marketers generate campaign briefs that pull live sentiment from posts, refining pitches on the fly.
—
Grab a moment to try these at grok.com. Poke around the new limits, share what you reckon on X, or sign up for updates to catch the next batch straight away.
