BAD AI: Lawyer Up
Dec 09, 2025
AI IS ACCELERATING - GET READY
Let's help you understand the WTF of AI and how it helps YOU and your organisation grow. Ready, set, agent.
This week:
😱 Bad AI: Lawyer Up
🔥 AI Offer: Get Free Brand Content
✨ AI Models: Kimi K2
💭 Creative Tools: Adobe Updates
💡 News, Trends & Insights
BAD AI: Lawyer Up
![]() |
Do you keep up with the Kardashians?
No, me neither. So why is Kim in our newsletter then?
She failed her law exams. Thanks for sharing that, got a point?
OK, I'll bite. Has it got something to do with AI? Yep.
Did she take advice directly from ChatGPT? Ooh, you're good.
Give us the juicy details. She mistook pro bono for an Irish advocacy group campaigning for U2's lead singer to be made the patron saint of AI*.
Was hallucination the issue? Well, I did watch 15 hours straight of KUWTK researching this section, so anything's possible tbf.
Ha ha. You funny this week. Thanks and with any luck this won't end up in Court(ney).
You can stop now.
*Not factually correct, as we went live.
AI CONTENT: Your Assets Repurposed For Free
Want to get more content out there for your brand?
Good. We can help.
Head over here: https://www.g3nr8.com/content-repurposing
We'll take a piece of your existing content and create visuals, social assets and posts with AI that are consistent and on-brand.
✅ Drop us a link to your content piece and details
✅ We'll create new assets using our AI agent system
✅ See how AI interprets your tone, visuals and messaging
✅ Experience what scaling (quality) content with AI actually feels like
It's the same system we deploy for our enterprise clients - you just get to see it work on your brand, firsthand.
👉 We've extended it - as we had a great response last week.
AI MODELS: KIMI K2
Moonshot AI has released a new open-source reasoning model, Kimi K2 Thinking, looking to match or exceed top frontier models - at a much lower cost.
Benchmark performance: Scored around 45% on the popular Humanity's Last Exam (HLE) benchmark, doing better than GPT-5 and Claude 4.5.
Autonomous workflows: Supports up to 300 sequential tool calls - meaning it can do extended reasoning and task chaining for longer without human input.
Architecture: Built on a Mixture-of-Experts (MoE) design with a 256K-token context window for long-form tasks.
Cost efficiency: Reportedly trained for under $5 million - and priced significantly lower than leading closed-source models.
What it means for businesses:
Kimi K2 shows that advanced reasoning isn't exclusive to the bigger Western labs - but businesses should weigh cost and capability against governance, data sovereignty, and compliance risks when adopting Chinese-developed AI.
Wider context:
China's open-source ecosystem is accelerating fast. As performance gaps close, trust, transparency, and geopolitical alignment will increasingly define which models global enterprises choose to build on.
AI CREATIVE: Adobe Updates

Adobe just held its 2025 MAX event and showcased turning Firefly into a fuller AI creative studio and embedding intelligent assistants directly inside Photoshop and Express.
Firefly 5: New model produces 4MP photoreal images with improved lighting, texture, and anatomy.
Unified AI plan: Combines Adobe, Google, OpenAI, and Runway models under one subscription.
Agentic assistants: Built into apps to automate edits, asset organisation, and repetitive creative tasks.
Partnerships: Deeper integrations with Google (Gemini, Veo) and YouTube for faster, cross-platform publishing.
What it means for businesses:
Adobe is streamlining production and brand consistency - giving teams faster asset turnaround and safer generative content. Whatever makes it more of a pain for enterprise creative ops already embedded in Adobe's ecosystem to move - that'll continue to be the focus.
Wider context:
Competition across the creative AI space is moving constantly. Canva's new OS, Figma's AI design features, and a wave of smaller players are all chasing the same vision - frictionless creativity powered by intelligent assistants.
For advice on the right models and tooling for your organisation, get in touch.
WHAT'S NEW - News, Insights & Trends
> UK school given AI-generated targets
> Word of the year is announced
> Google warns about AI malware
FAQ: What is Kimi K2 and how does it compare to other AI models?
Kimi K2 is an open-source reasoning model from Moonshot AI that scored around 45% on the Humanity's Last Exam benchmark, outperforming GPT-5 and Claude 4.5. It uses a Mixture-of-Experts architecture with a 256K-token context window and supports up to 300 sequential tool calls for extended autonomous workflows.
FAQ: How much did Kimi K2 cost to train?
Kimi K2 was reportedly trained for under $5 million, making it significantly more cost-efficient than leading closed-source models while delivering competitive benchmark performance.
FAQ: What are the key features of Adobe Firefly 5?
Adobe Firefly 5 produces 4MP photoreal images with improved lighting, texture, and anatomy. It's part of Adobe's new unified AI plan that combines models from Adobe, Google, OpenAI, and Runway under one subscription, with agentic assistants built into apps to automate edits and repetitive creative tasks.
FAQ: What should businesses consider when adopting Chinese-developed AI models?
Businesses should weigh cost and capability against governance, data sovereignty, and compliance risks when adopting Chinese-developed AI. As performance gaps close between Chinese and Western models, trust, transparency, and geopolitical alignment will increasingly define which models global enterprises choose to build on.
FAQ: What is Adobe's unified AI plan?
Adobe's unified AI plan combines models from Adobe, Google (Gemini, Veo), OpenAI, and Runway under one subscription. It includes deeper integrations with YouTube for faster cross-platform publishing and agentic assistants built directly into Photoshop and Express.
FAQ: How is AI changing the creative tools landscape?
Competition across the creative AI space is intensifying. Adobe is streamlining production and brand consistency, while Canva's new OS, Figma's AI design features, and smaller players are all chasing frictionless creativity powered by intelligent assistants. The focus is on faster asset turnaround and safer generative content.
FAQ: What does Kimi K2's 300 sequential tool calls capability mean?
Kimi K2's support for up to 300 sequential tool calls means it can perform extended reasoning and task chaining for longer periods without human input, enabling more sophisticated autonomous workflows compared to models with lower tool call limits.
G3NR8 is an impact consultancy for the exponential age. We help organisations implement AI for maximum returns on investment.
Work with us to:
- Rapidly improve your sales and marketing function with an AI ecosystem - from horizon scanning to content repurposing and active campaign planning.
- Help each employee find information 60% faster with advanced AI knowledge bases - secure, and based on your business logic and data.
- Create valuable MVPs that don't sit in silos and work immediately to generate productivity gains and ROI in weeks, not months.
- Train your teams - we've trained 150+ large organisations in advanced generative principles and implementation.
Drop us a line for a chat about productivity gains, cost savings, growth and how AI can scale your business.
Stay Ahead with The Ultimate AI Newsletter
Subscribe for unique AI insights and strategies that redefine business and innovation. Plus, get VIP access to a curated selection of "bad AI" - because sometimes, learning what not to do is just as valuable..
