Top 15 AI News of 2025 : GPT-5 Backlash, Open Models Rise & AI Regulation Reality (Part 2/3)

Welcome to Part 2 of our Top 15 AI News of 2025, where we cover news #10 to #6 — the stories that shaped how AI is built, regulated, and adopted this year.

In this episode, we break down:
The real state of AI regulation in the US and globally
Why GPT-5 triggered massive backlash — and what OpenAI reversed
The rise of open-weight and open-source AI models
Why benchmarks are flattening and no longer reflect real progress
The creative breakthrough of NanoBanana and continuity in AI image editing

This is not hype coverage. We analyse what these events actually mean for businesses, professionals, creators, and governments using AI at scale.

This is Part 2 of 3 - Part 3 will cover the Top 5 AI News of 2025.

Timestamps
00:00 – Introduction: Top 15 AI News (Part 2)
00:06 – #10: US AI Regulation & the rollback of safety rules
00:51 – “Winning the race”: what US regulation is really about
01:29 – Global AI regulation map: who regulates, who doesn’t
01:51 – #9: The release of GPT-5
02:30 – Why GPT-5 wasn’t a real breakthrough
02:49 – Backlash: retiring GPT-4 and user frustration
03:06 – The router problem: AI deciding for the user
03:50 – Trust, cost, and loss of control
04:18 – OpenAI reverses course: restoring models & selectors
05:07 – B2C vs B2B: why OpenAI listens to users
05:33 – #8: The rise of open-weight AI models
06:07 – DeepSeek R1 and why open models matter for enterprises
06:57 – OpenAI releases GPT-OSS under Apache 2.0
07:33 – Running AI locally: data protection implications
08:02 – #7: AI benchmarks and performance plateau
08:53 – Why progress looks flat in 2025
09:49 – When benchmarks stop measuring real value
10:23 – Why new benchmarks are needed
10:46 – #6: NanoBanana and AI image continuity
11:31 – Why continuity is a major creative breakthrough
12:37 – NanoBanana vs Quan Image Edit
13:58 – What this means for designers and creators
14:27 – What’s coming next in Part 3 (Top 5 AI News)





































AI news 2025, GPT-5 backlash, AI regulation US, open-weight AI models, open source AI 2025, DeepSeek R1, GPT-OSS, AI benchmarks plateau, AI performance limits, NanoBanana AI, AI image continuity, generative AI trends, enterprise AI adoption, future of AI