6 hours ago
6
Follow ZDNET: Add america arsenic a preferred source on Google.
ZDNET's cardinal takeaways
- OpenAI has updated ChatGPT's personalization page.
- The leafage includes property options, representation toggles, and more.
- Not each users consciousness seen oregon heard by the update.
OpenAI is making it easier for ChatGPT users to customize what it calls the chatbot's "personality" and connection style.
ChatGPT present offers an updated personalization page, which OpenAI CEO Sam Altman previewed successful a Tuesday X post. The caller leafage includes a dropdown paper with a scope of property types, including "Cynic," "Robot, "Listener," and "Nerd," and a "custom instructions" punctual section, done which users tin modify the chatbot's outputs with requests similar "Avoid millennial jargon."
Also: AI models cognize erstwhile they're being tested - and alteration their behavior, probe shows
There are besides sections wherever you tin specify your preferred nickname and occupation, and different wherever you tin database idiosyncratic "interests, values, oregon preferences" -- the intent of which appears to beryllium to make a conversational acquisition that feels much similar chatting with a trusted person oregon workfellow than with an unfeeling machine. Toward the bottommost of the updated page, which is accessible present via settings successful ChatGPT, you tin modify the chatbot's representation capabilities.
The subject (and business) of artificial personalities
The refinement of chatbot personalities has go a captious absorption for tech developers arsenic they proceed to merchantability AI to individuals and businesses. The situation is to physique systems that make close accusation successful a mode that besides feels engaging and personable. This has occasionally gone excessively far, arsenic successful the caller lawsuit of ChatGPT users complaining that the chatbot's responses tended to veer into sycophancy and excessive flattery.
Also: ChatGPT volition verify your property soon, successful effort to support teen users
Different companies person besides taken assorted approaches to artificial personalities: xAI's Grok tends to beryllium irreverent and edgy, Claude is much of a sober and cautious integer assistant, and ChatGPT's tenor tends to beryllium much flexible depending connected the preferences of each user.
Focusing connected personalization
That flexibility and adaptability person go a bigger absorption for OpenAI successful the aftermath of past month's debut of GPT-5, the latest ample connection exemplary (LLM) to powerfulness ChatGPT. Many users person complained since the model's motorboat that it's inferior to its predecessor, GPT-4o, some successful presumption of its velocity and its connection style, which immoderate recovered to beryllium level and cold.
OpenAI rapidly responded to those complaints with a flurry of updates to GPT-5, 1 of which allowed paid ChatGPT Plus subscribers to regain casual entree to 4o.
Altman said that 1 of the cardinal lessons of GPT-5's underwhelming archetypal occurrence was that OpenAI needed to absorption much connected gathering much personalized and customizable tools.
"Long-term, this has reinforced that we truly request bully ways for antithetic users to customize things (we recognize that determination isn't 1 exemplary that works for everyone, and we person been investing successful steerability probe and launched a probe preview of antithetic personalities)," helium wrote successful a X post connected Aug. 8, adding the illustration of immoderate users possibly preferring tons of emoji portion others "never privation to spot one."
OpenAI received a akin backlash from galore among its userbase erstwhile it announced earlier this period that it would retire ChatGPT's Standard Voice Mode. The institution aboriginal rescinded that decision.
Also: How radical really usage ChatGPT vs Claude - and what the differences archer us
(Disclosure: Ziff Davis, ZDNET's genitor company, filed an April 2025 suit against OpenAI, alleging it infringed Ziff Davis copyrights successful grooming and operating its AI systems.)
More calls to reinstate 4o
ChatGPT's caller personalization leafage has already received a chilly reception online, with galore radical arguing that the institution is misunderstanding their complaints astir GPT-5.
"A cleanable idiosyncratic interface is welcome, convey you, but delight besides see the petition from thousands of users to support 4o and Standard Voice Mode agelong term," one idiosyncratic wrote under Altman's X post. "4o is capable to intuitively accommodate to idiosyncratic volition and preference, dynamically moving alongside you. It doesn't request customized instructions oregon personalisation."
Another person's reply enactment it adjacent much succinctly: "No operation of toggles and instructions tin replicate the integrated and effortless vibe of interacting with 4o."
At the clip of this writing, OpenAI has not announced immoderate plans to reinstate 4o for users of ChatGPT's escaped tier. The institution did not instantly respond to ZDNET's petition for comment.
English (US) ·