美女免费一级视频在线观看

OpenAI launched ChatGPT-5 on Thursday (7 August) last week, which sent out a “warning shot” across adland, according to creatives.

Ross Taylor, group executive creative director at Havas Play, told Campaign: “The launch of ChatGPT-5 isn’t just an upgrade, it’s a warning shot for the advertising industry. We’re no longer talking about tools that assist us; we’re talking about tools that simulate us.”

The upgraded model was designed with “expert level” reasoning and systems that decide when to answer quickly or think longer for more complex questions.

OpenAI chief executive Sam Altman compared GPT-5’s intelligence to PhD-level expertise, where GPT-4 was a college student and GPT-3 a high-school student. 

There is no doubt that GPT’s updates can help users and advertisers “move quicker and, more importantly, think weirder”, said Taylor. “It can remix culture in ways we hadn’t imagined, challenge tropes, sparking ideas we might never have found alone.”

But he questioned: “Do we really need writers or just people that can prompt and review? As AI gets closer to being more human-like, it pushes us to ask: what’s uniquely human about what we do? And that question will only lead to work with more humanity at its heart – a good thing in my book!”

GPT-5 has also fuelled existing concerns about ownership with AI-created content. OpenAI claimed the updated model supports “more human-like writing”, with improved tone, rhythm and structure in poetry, essays and storytelling, as well as “more reliable” professional writing in memos, reports and emails.

R/GA’s EMEA chief creative officer Nick Pringle warned: “We’re tiptoeing into murky waters.”

As GPT and other AI systems improve their ability to replicate human-made content, he questioned who owns the “soul” of the output. 

Pringle said: “With AI remixing everything from eulogies to screenplays, the lines around consent are blurring fast. Brands and platforms will need to bake in permissions at the model level, or risk a future full of legal headaches and PR disasters when ‘inspired by’ turns into ‘stolen from’.”

Getty Images’ chief product officer Grant Farhall, meanwhile, expressed concern to Campaign about the commercial risk that comes with how AI models are being trained as they “reshape the creative process”.

He said: “Many AI tools are still being trained on scraped content that violates the rights of IP holders, and can therefore create outputs that contain elements protected by copyright. Brands should also consider whether creators are being fairly compensated for their works being trained.

“Artists need to be able to thrive in this environment, and that only happens if their rights are respected and they are compensated for the work they create, including when it is used as training data. Otherwise, eventually, that work will simply go away. That’s bad for everyone.”

Pringle shared that GPT-5 “continues the shifts in gears from creator-as-craftsperson to creator-as-conductor”. 

He explained that creativity is “forever pedestalled as mystical and deeply human”, but as it becomes increasingly replicable and “frighteningly scalable”, there will be a rise of content that is “emotionally compelling enough to pass as human-made”. 

Pringle added: “GPT-5 means we’ll be drowning in fast, polished, technically brilliant [content].” 

However, he also argued that with the increase of machine-made content, audiences may start craving the “messy, imperfect, unmistakably human stuff” that feels like it has “soul”.

Pringle said: “Perhaps human-made content will become a niche, like vinyl… only for those who know or care about the difference.”

Getty Images research showed that people value authenticity and transparency, Farhall claimed, especially in visual storytelling. “The challenge for brands will be harnessing AI’s capabilities without losing the human nuance and trust that makes content truly resonate,” he said.

In addition, Taylor and Pringle believe that the “feel” of work will become more important in light of technology such as ChatGPT-5 being used in adland.

Taylor said: “This isn’t about fighting the tech. It’s about proving that human insight and emotion still matter. Real creativity isn’t predictable; it’s messy, experienced, lived. GPT-5 can play the game, but only we can truly feel it.”

Pringle, meanwhile, believes that emotional realism will start to trump source legitimacy, as people’s connection with AI deepens. He said: “In short, we’ll care less who wrote it, and more how it makes us feel. Authenticity, ironically, will become an output, not an origin.” 

Alongside concerns from adland over the abilities of GPT-5, its launch was also met with criticism from some of its 100 million monthly active users, as it imposed limits on the amount of questions users could ask in a day. Others missed the previous model, GPT-4o – a more advanced version of GPT-4 – which was initially no longer accessible after GPT-5 launched.

Altman swiftly brought back GPT-4o to be used within GPT-5, claiming on X: “We expected some bumpiness as we roll out so many things at once. But it was a little more bumpy than we hoped for.”

OpenAI appointed Omnicom Media Group’s PHD to its global media account last week.

OpenAI has been contacted for comment.

This article originally appeared on Campaign UK.