28.1 C
Delhi
Monday, March 2, 2026

OpenAI Strengthens Sora 2 Guardrails After Bryan Cranston Alarm

OpenAI Strengthens Sora 2 Guardrails After Bryan Cranston Raises Alarm

OpenAI has significantly enhanced protections against unauthorized content generation in Sora 2 following concerns raised by actor Bryan Cranston about the AI’s ability to replicate celebrity likenesses and copyrighted material without permission.

Key Takeaways

  • OpenAI strengthened Sora 2 guardrails after Bryan Cranston flagged unauthorized videos
  • New protections prevent replication of real people’s likenesses without explicit consent
  • Joint effort with SAG-AFTRA and talent agencies to ensure performer protections
  • System now blocks copyrighted characters like Spongebob, Pikachu, and Mario

The Cranston Catalyst

During Sora 2’s September 30 launch, OpenAI initially promised to prohibit unauthorized likeness replication through a “cameo” opt-in feature. However, videos featuring Bryan Cranston as Walter White from “Breaking Bad” quickly appeared alongside AI-generated content of other celebrities including Michael Jackson and Ronald McDonald.

Cranston raised the issue with SAG-AFTRA, triggering a collaborative effort between the union, OpenAI, and talent agencies. “I felt deeply concerned not just for myself, but for all performers whose work and identity can be misused in this way,” Cranston stated.

Enhanced Protection Measures

OpenAI CEO Sam Altman confirmed the company is “deeply committed to protecting performers from the misappropriation of their voice and likeness.” The strengthened guardrails now prevent generation of content resembling third-party likenesses or copyrighted material.

Previously, Sora 2 allowed generation of copyrighted material unless rights holders opted out. Now, such attempts return error messages stating prompts “may violate our guardrails.”

Hollywood’s AI Dilemma

The development comes as Hollywood professionals grapple with AI advancements. While many secretly use AI tools, tensions remain high over potential misuse of likenesses and intellectual property.

Talent agency CAA had earlier criticized OpenAI for “exposing our clients and their intellectual property to significant risk” by allowing generation of copyrighted characters.

Legislative Support and Industry Response

The joint statement expressed support for the NO FAKES Act, legislation introduced in April that would hold entities liable for producing or hosting unauthorized deepfakes. “We were an early supporter of the NO FAKES Act and will always stand behind the rights of performers,” Altman wrote.

SAG-AFTRA President Sean Astin commended the resolution, stating: “Bryan Cranston is one of countless performers whose voice and likeness are in danger of massive misappropriation by replication technology. This particular case has a positive resolution.”

Cranston expressed gratitude for OpenAI’s policy improvements, hoping companies “respect our personal and professional right to manage replication of our voice and likeness.”

Latest

Sam Altman reveals real reason why OpenAI rushed to partner with US Military after Trump banned Anthropic

OpenAI executives have given more information regarding the AI startup’s contract with the US Department of Defense after facing backlash online. The Sam Altm

After Donald Trump banned Anthropic, US Military used Claude in Iran strikes: Here is what changed

The US Military reportedly used Anthropic’s Claude AI model during its strikes on Iran. The attack on Iran came just a day after US President Donald Trump ins

SIM binding rules go live starting March 1: These WhatsApp, Telegram, Signal and other messaging app users to be impacted

Tech News News: Starting March 1, messaging apps like WhatsApp, Telegram, Signal and others must comply with the Department of Telecommunications' SIM-binding r

More than one year after DeepSeek’s R1 wiped nearly $600 billion off Nvidia market value in single day, Chinese startup planning another launch

Tech News News: DeepSeek, the Chinese AI startup that wiped nearly $600 billion off Nvidia’s market value in a single day with launch of its R1 model, is repo

Nothing Phone 4a and 4a Pro launching on 5 March: Design, expected specs and more

Nothing is set to launch its Phone 4 (a) series on 5 March. The launch event is also likely to see the unveling of new Headphone (a) with bold colors and long b

Topics

Taliban attacks Pak’s Nur Khan base in latest escalation of cross border conflict

Taliban forces reportedly launched armed drone strikes targeting Pakistan’s Command and Control Centre at Nur Khan Air Base in Rawalpindi. Taliban forces carr

Satellite images show damage across Iranian military sites after US-Israel strikes

Fresh satellite imagery shows visible damage to air, drone and naval facilities near Iran’s Konarak region amid escalating regional tensions. The visuals offe

Sensex down 1,000 points: Why is the stock market falling today?

The S&P BSE Sensex fell sharply in early trade, and the NSE Nifty50 also slipped more than 1%, as investors reacted to the fast-changing situation between the U

Qatar, UAE, Syria, Oman: Full list of places that saw attacks amid US-Iran conflict

The Middle East is engulfed in conflict as Iran retaliates against US-Israeli strikes, launching missile and drone attacks across multiple countries. 

AIIMS-trained neurologist warns against repeatedly using reheated cooking oils: ‘Risk of cancer increases manifold…’

Reusing cooking oil is a common practice in many households, but does the money it saves outweigh the health risks? Dr Sehrawat explains the health risks.

Quote of the day by Jon Bon Jovi: ‘You better stand tall when they’re calling you out, don’t bend, don’t break…’

On his birthday, we look back at one of Jon Bon Jovi's most influential quotes, which highlights the importance of standing tall in the face of criticism.

Satellite images show black smoke over Dubai as Iran continues to fire missiles, drones

Iran-US war: Dubai's skyline has dramatically changed after Iranian attacks, with smoke visible in satellite images.

Sam Altman reveals real reason why OpenAI rushed to partner with US Military after Trump banned Anthropic

OpenAI executives have given more information regarding the AI startup’s contract with the US Department of Defense after facing backlash online. The Sam Altm
spot_img

Related Articles

Popular Categories

spot_imgspot_img